Apparatus and method for detecting behavior change on basis of digital behavior

Abstract

The present invention relates to an apparatus for detecting a behavior change on the basis of a digital behavior, the apparatus comprising: an acquisition unit for acquiring, from a user terminal, digital behavior data of a user; a reference setting unit for setting reference behavior characteristic information of the user through analysis of the digital behavior data; and a change search unit for searching additional digital behavior data for a change in a behavior different from a usual behavior of the user through analysis of the additional digital behavior data additionally acquired after the reference behavior characteristic information is set.

Description

Digital behavior-based behavior change detection device and method

[0001] The present invention relates to a device and method for detecting behavioral changes based on digital behavior. In particular, the present invention is not intended for performing medical diagnoses, but relates to a technology for monitoring a user's general wellness state and screening signals of behavioral changes.

[0002] Changes in a user's cognitive or physical functions can lead to various daily difficulties, and these changes are often noticed first by caregivers, such as family members, rather than by the individual themselves.

[0003] However, current welfare systems generally only initiate intervention after a user's condition (such as mild cognitive impairment (MCI) or dementia) has been clearly identified. Consequently, a system for delivering information on daily life responses—specifically, based on small, repetitive behavioral changes in daily life—is not yet sufficiently established for the stage prior to such clear identification. In other words, conventional technology has not been available to help address changes in a user's behavior more quickly by detecting and notifying them of small, repetitive behavioral changes (information on daily life changes / life responses) before an abnormal condition is clearly detected.

[0004] If a user's cognitive or physical decline persists for a long time, the emotional, psychological, and economic burdens borne by families and society can be significant. For example, when a single individual experiences cognitive decline, there are many cases where more than 10 people around them suffer complex burdens, including emotional fatigue, exhaustion of caregiving time, and reduced work life, in addition to medical expenses. Therefore, there is an urgent need for the development of technology that can alleviate these burdens in advance, before they fully materialize.

[0005] In addition, if a user's cognitive decline persists for a long period or is left untreated, it is highly likely to lead to a lonely death, and there are statistics showing that it actually takes an average of 17.6 days (family), 29.7 days (neighbors / landlords), and 67.8 days (welfare officials) to discover the body after death.

[0006] Accordingly, there is an urgent need for the development of technology that analyzes users' digital behavior data (daily data) naturally generated by their use of user devices in daily life, observes (searches for) changes in behavior (abnormal behavior) that differ from the norm caused by cognitive decline, and enables the rapid detection and response to such behavioral changes based on this analysis.

[0007] The technology forming the background of the present invention is disclosed in Korean Published Patent Application No. 10-2020-0104758 ("Method and apparatus for determining a dangerous situation and managing user safety"). However, the aforementioned prior art is merely a technology that determines whether a dangerous situation exists based on changes in the user's biosignals and, if the user is in a dangerous situation, determines the level of danger and notifies a guardian terminal. It does not disclose at all a technology that detects small, repetitive behavioral changes occurring in the user during daily life at an early stage through the analysis of data from the user's smartphone (user terminal) usage, thereby inducing the user to pay attention to changes in the user's behavior more quickly.

[0008] The present invention aims to solve the problems of the aforementioned prior art by providing a digital behavior-based behavior change detection device and method that analyzes a user's digital behavior data (daily data) naturally generated by the user's use of a user terminal in daily life, observes (searches for) changes in behavior (abnormal behavior) that differ from the usual due to cognitive decline or the like, and induces rapid detection and response to such behavior changes based on this.

[0009] The present invention aims to solve the problems of the aforementioned prior art by providing a digital behavior-based behavior change detection device and method that can help users take notice of changes in their behavior more quickly by detecting and notifying them of information regarding small, repetitive behavioral changes occurring in daily life (information regarding changes in daily life / life reactions) before an abnormal state is clearly detected in the user, and also provide an opportunity to initiate gentler conversation and interest with the user by quietly conveying to pre-set notification recipients information about changes (behavioral changes) that naturally occur in the flow of daily life before the emotional, psychological, and economic burdens caused by the deterioration of cognitive or physical functions in the user begin in earnest.

[0010] The present invention aims to provide a digital behavior-based behavior change detection device and method that is not intended for medical judgment or disease diagnosis, but merely to help enable changes in a user's daily life to be shared earlier and more naturally within a warm, family-centered care system.

[0011] This invention is not a technology intended for medical diagnosis, definitive determination of disease, or clinical prescription, but rather relates to a technology for monitoring a general wellness state and screening signals of behavioral changes based on a user's daily digital behavior patterns. The purpose of this invention is not to determine the presence of a specific disease or to replace medical decisions, but rather to provide reference information by detecting statistical change trends appearing in a user's behavioral characteristics at an early stage.

[0012] However, the technical problems that the embodiments of the present invention aim to solve are not limited to the technical problems described above, and other technical problems may exist.

[0013] As a technical means for achieving the above-mentioned technical task, a digital behavior-based behavior change detection device according to one embodiment of the present invention may include: an acquisition unit that acquires digital behavior data of a user from a user terminal; a standard setting unit that sets standard behavior characteristic information of a user through analysis of the digital behavior data; and a change detection unit that searches for a change in behavior different from the user's usual behavior within the additional digital behavior data through analysis of additional digital behavior data acquired after the setting of the standard behavior characteristic information.

[0014] In addition, the change detection unit can detect a change in the behavior by comparing the current behavior characteristic information derived from the additional digital behavior data with the reference behavior characteristic information.

[0015] In addition, the change detection unit can not only determine whether there is a change by comparing the current behavioral characteristic information with reference behavioral characteristic information, but can also detect abnormal changes by comparing with the average behavioral characteristics of multiple users who are of the same age group, gender, or similar usage characteristics as the user (i.e., group reference behavioral characteristic information).

[0016] In addition, the digital behavior data includes user typing data, and the standard setting unit can derive a typing pattern regarding at least one of typing speed, rhythm, typing position, number of typos, and number of deletion inputs through analysis of the typing data and set it as typing-related standard behavior characteristic information.

[0017] In addition, the digital behavior data includes the user's voice data, and the standard setting unit can derive a voice pattern regarding at least one of the user's speech rhythm, prosodic features, and disfluency features through the analysis of the voice data and set it as voice-related standard behavior characteristic information.

[0018] In addition, the digital behavior data includes user activity data, and the standard setting unit can derive at least one of an app usage pattern, an inactive state pattern, and a location pattern through the analysis of the activity data and set it as activity-related standard behavior characteristic information.

[0019] In addition, a digital behavior-based behavior change detection device according to one embodiment of the present invention may further include a notification management unit that transmits notification information indicating that there is a change in the user's behavior when the change in behavior is detected. Here, the notification information may refer to notification information intended to induce emotional interest from family (or guardian), etc., rather than inducing medical measures.

[0020] In addition, a digital behavior-based behavior change detection device according to one embodiment of the present invention further includes a control unit that provides today's sentence to a user terminal, and the change search unit can search for the change by further considering analysis information of the sentence input data when user sentence input data corresponding to today's sentence, generated by the user inputting today's sentence, is collected from the user terminal.

[0021] In addition, the control unit can provide a reward to the user terminal when the user inputs the sentence input data.

[0022] Meanwhile, a method for detecting a change in behavior based on digital behavior using a digital behavior-based behavior change detection device according to one embodiment of the present invention may include: a step of acquiring digital behavior data of a user from a user terminal in an acquisition unit; a step of setting standard behavior characteristic information of a user through analysis of the digital behavior data in a standard setting unit; and a step of searching for a change in behavior that is different from the user's usual behavior within the additional digital behavior data through analysis of additional digital behavior data additionally acquired after the setting of the standard behavior characteristic information in a change search unit.

[0023] The above-described means for solving the problem are merely exemplary and should not be interpreted as intended to limit the invention. In addition to the exemplary embodiments described above, additional embodiments may exist in the drawings and the detailed description of the invention.

[0024] According to the means for solving the problem of the present invention described above, by providing a device and method for detecting behavioral changes based on digital behavior, the digital behavioral data (daily data) of a user that naturally occurs (is generated) as a user uses a user terminal in daily life is analyzed to observe (search) for behavioral changes (abnormal behavior) that differ from the usual behavior appearing in the user due to cognitive decline, etc., and based on this, rapid detection and response to such behavioral changes can be induced.

[0025] According to the means for solving the problem of the present invention described above, by providing a digital behavior-based behavior change detection device and method, information regarding small, repetitive behavioral changes occurring in daily life (information regarding changes in daily life / life reactions) is detected and provided before an abnormal state is clearly detected by the user, thereby helping the user to pay attention to changes in behavior more quickly.

[0026] According to the means for solving the problem of the present invention described above, by providing a device and method for detecting behavioral changes based on digital behavior, changes (behavioral changes) that naturally occur in the user's daily routine can be quietly conveyed to pre-set notification recipients before the emotional, psychological, or economic burden caused by a decline in cognitive or physical functions of the user intensifies; thereby, an opportunity can be provided to initiate gentler conversation and interest toward the user.

[0027] According to the means for solving the problem of the present invention described above, by providing a device and method for detecting behavioral changes based on digital behavior, it is possible to help share changes in the user's daily life earlier and more naturally within a warm, family-centered care system, rather than for the purpose of medical judgment or disease diagnosis.

[0028] However, the effects obtainable from the present invention are not limited to those described above, and other effects may exist.

[0029] FIG. 1 is a diagram showing the schematic configuration of a digital behavior-based behavior change detection device according to one embodiment of the present invention.

[0030] FIG. 2 is a diagram schematically illustrating the processing process for typing data in a digital behavior-based behavior change detection device according to one embodiment of the present invention.

[0031] FIG. 3 is an operation flowchart of a digital behavior-based behavior change detection method according to an embodiment of the present invention.

[0032] Embodiments of the present invention are described below with reference to the attached drawings so that those skilled in the art can easily implement the invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly explain the present invention in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification are denoted by similar reference numerals.

[0033] Throughout the specification of the present invention, when a part is described as being "connected" to another part, this includes not only cases where they are "directly connected," but also cases where they are "electrically connected" or "indirectly connected" with other elements interposed between them.

[0034] Throughout the specification of the present invention, when a member is described as being located "on," "on the upper," "on the top," "under," "on the lower," or "on the bottom" of another member, this includes not only cases where the member is in contact with the other member, but also cases where another member exists between the two members.

[0035] Throughout the specification of the present invention, when a part is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0036] Throughout the specification of the present invention, some of the operations or functions described as being performed by a terminal, device, or device may instead be performed by a server connected to said terminal, device, or device. Likewise, some of the operations or functions described as being performed by a server may also be performed by a terminal, device, or device connected to said server.

[0037] Throughout the specification of the present invention, the term "at least one" may be defined as a term including both singular and plural forms, and it will be obvious that even if the term "at least one" is not present, each component may exist in a singular or plural form and may mean singular or plural. Furthermore, whether each component is provided in a singular or plural form may be changed according to the embodiment.

[0038] FIG. 1 is a diagram showing the schematic configuration of a digital behavior-based behavior change detection device (10) according to one embodiment of the present invention.

[0039] For convenience of explanation, the digital behavior-based behavior change detection device (10) according to one embodiment of the present invention will be referred to as the device (10) below. In addition, details shown (described) in the drawings of the present invention, including FIG. 1, may be applied equally to the description of the device (10) below, even if such details are omitted.

[0040] Furthermore, the terms “detection” or “judgment” in this specification shall be interpreted to mean the confirmation of a signal of statistical change and do not imply a medical diagnosis, confirmation of a disease, or clinical evaluation. Additionally, the expressions “judgment of change” or “abnormal sign” in this specification refer to changes in behavioral patterns from a wellness perspective and do not include a definitive judgment regarding the presence of a medical disease.

[0041] Referring to FIG. 1, the device (10) may mean a digital behavior-based behavior change detection device that can observe (explore) a change in behavior (behavior change) different from the user's usual behavior by collecting and analyzing the user's digital behavior data.

[0042] The device (10) may be a device or server that provides at least one of a webpage (homepage), program, application (app), service, and platform related to the detection of behavioral changes based on digital behavior. At this time, the webpage (homepage), program, application (app), service, and platform related to the detection of behavioral changes based on digital behavior provided by the device (10) shall be referred to as the webpage (this homepage), the program, the app, the service, and the platform, respectively, for convenience of explanation below. The service name of the service may be Onki, etc., as an example, but is not limited thereto.

[0043] The device (10) may be a device that can be implemented, for example, in the form of a program or application (app) installed on a user terminal (20). Alternatively, the method for detecting behavioral changes based on digital behavior provided through the device (10) may be implemented in the form of a program or application and provided to the user through the user terminal (20). However, it is not limited thereto, and as another example, the device (10) may be provided in the form of a server capable of transmitting and receiving data with the user terminal (20). The device (10) provided in the form of a server may control the operation (for example, screen display operation, etc.) of the user terminal (20) connected to the app or program provided by the device (10).

[0044] When the device (10) is provided in the form of a server, the digital behavior-based behavior change detection system (hereinafter referred to as "the system") according to one embodiment of the present invention may include the device (10), a user terminal (20), and an administrator terminal (not shown). In this case, the device (10) may be linked with the user terminal (20) and the administrator terminal (not shown), respectively, via a network to transmit and receive data. Additionally, the device (10) provided in the form of a server may be referred to by other terms, such as server, the server, etc.

[0045] The present system may include a user terminal (20) that generates digital behavior data, and a device (10) that receives the digital behavior data or feature information generated based on the digital behavior data from the user terminal (20) and detects a change in behavior.

[0046] Here, the above feature information may refer to statistical / numerical characteristic values ​​derived by analyzing source digital behavior data collected from a user within a user terminal (20). For example, the above feature information may consist of an average value, standard deviation, coefficient of variation, frequency value, ratio value, time series trend value, slope of change, rate of change, acceleration indicator, or a combination thereof, and may not include data that can directly identify an individual, such as source input content, original voice waveform, or location path.

[0047] Additionally, the above feature information may be statistical summary information in a de-identified state so that individual user identification is impossible, and may include the result of a feature extraction operation performed within the user terminal (20) or a vectorized feature representation for model input. Accordingly, the above feature information may refer to an abstract data set that numerically represents behavioral characteristics, rather than the content of the source data itself.

[0048] In the present invention, the results of detecting (searching, screening) behavioral changes are not interpreted as medical judgments or clinical diagnoses, and the information provided to users or third parties may be utilized as reference indicators for the purposes of health management, wellness monitoring, or lifestyle pattern management. Furthermore, the detection results merely present the statistical possibility of the presence of a risk signal and do not directly determine the existence of a disease, the degree of progression, or the necessity of treatment.

[0049] The device (10), which is a server, can manage long-term trends of behavioral characteristics for each user and select change signals, but the device (10) is not configured as a diagnostic system or clinical decision support system of a medical institution and does not include a direct diagnostic linkage function with a medical database as an essential component. In addition, the system is not intended to be combined with medical imaging, biosignal-based diagnostic devices, or clinical testing equipment, and can be implemented as a wellness screening platform based on daily digital behavior data.

[0050] Users of the device (10) may include users and administrators. Users may use the service with or without installing the program or app using their own terminals. Additionally, users may use the service with or without registering (logging in) by accessing the program or app using their own terminals.

[0051] The administrator terminal (not shown) may refer to a terminal possessed by the administrator. The administrator may be a person who develops and distributes the device (10) and operates and manages the device (10).

[0052] The user terminal (20) may refer to a terminal possessed by the user. The user may refer to a subject (detection subject) for which digital behavior-based behavior change detection (search) is performed by the device (10). The user may be a person who uses the device (10), may be referred to by other terms such as customer or user, and may be anyone regardless of gender or age.

[0053] After installing the app on the user terminal (20), the user can, for example, access the app and set the activation status of the third-party keyboard to an enabled state, and through the setting to the enabled state, the third-party keyboard can be installed on the user terminal (20). The user can perform typing input on the user terminal (20) using the installed third-party keyboard.

[0054] Here, the third-party keyboard may refer to a keyboard (keyboard app) developed (produced) and distributed by an administrator, and may refer to a keyboard directly developed by an administrator for the purpose of acquiring (collecting) user typing data as the user performs typing input on the user terminal (20).

[0055] In the present invention, the terminal (i.e., the user terminal (20) and the administrator terminal (not shown), each) may include, for example, all types of wired and wireless communication devices such as PCS (Personal Communication System), GSM (Global System for Mobile communication), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (WCode Division Multiple Access), Wibro (Wireless Broadband Internet) terminal, smartphone, smartpad, tablet PC, laptop, wearable device, desktop PC, etc., but is not limited thereto.

[0056] In addition, the network described above may include, for example, a 3GPP (3rd Generation Partnership Project) network, an LTE (Long Term Evolution) network, a WIMAX (World Interoperability for Microwave Access) network, the Internet, a LAN (Local Area Network), a Wireless LAN (Wireless Local Area Network), a WAN (Wide Area Network), a PAN (Personal Area Network), a Bluetooth network, a NFC (Near Field Communication) network, a satellite broadcasting network, an analog broadcasting network, a DMB (Digital Multimedia Broadcasting) network, but is not limited thereto and may include various wired / wireless communication networks.

[0057] In one example of the present invention, the case in which one user terminal (20) is included within the present system (not shown) is described as an example, but this is merely an example to aid in understanding the present invention and is not limited thereto. Preferably, the present system (not shown) may include multiple user terminals possessed by each of the multiple users, and in describing the present invention below, the description of the user terminal (20) may be applied equally to the description of each of the multiple user terminals, even if the description is omitted below. The user of the user terminal (20) may be any one of the multiple users using the present device (10) and may be referred to as the first user, etc. Furthermore, in describing the present invention, the description of the user (first user) may be applied equally to the description of each of the multiple users. In addition, among the multiple users using the present device (10), the remaining users excluding the first user may be referred to by terms such as other users.

[0058] In the following description of the device (10), the device (10) is described using an example in which it is provided (implemented) in a form included (installed) within a user terminal (20), rather than in the form of a server. Referring to FIG. 1, the device (10) may be implemented in the form of the program or the app and included (installed) within the user terminal (20).

[0059] In addition, in describing the device (10) below, FIG. 1 illustrates that all (entirely) the components within the device (10) (i.e., keyboard setting unit, acquisition unit, database unit, standard setting unit, change detection unit, notification management unit, control unit) are provided within the user terminal (20), but this is merely an example to aid in understanding the invention and is not limited thereto. As another example, some components (e.g., notification management unit) among all the components within the device (10) may be provided within the server (the server). However, among the components within the device (10), some components (e.g., acquisition unit, database unit, standard setting unit, and change detection unit) particularly related to the acquisition and analysis of digital behavior data may be provided within the user terminal (20) so that they can be processed locally (within the user terminal) for the protection of the user's personal information (privacy protection).

[0060] In addition, in describing the device (10) below, the user terminal (20) including the device (10) will be described as a smartphone having portability (mobility) for the convenience of explanation. A detailed description of the device (10) provided within the user terminal (20) is as follows.

[0061] Referring again to FIG. 1, the device (10) may include a keyboard setting unit (11), an acquisition unit (12), a database unit (13), a standard setting unit (14), a change search unit (15), a notification management unit (16), and a control unit (17).

[0062] The keyboard setting unit (11) can enable a third-party keyboard developed by an administrator to be installed in the user terminal (20). For example, the keyboard setting unit (11) can enable the user to install the app on the user terminal (20) and then enable the setting for whether to enable the third-party keyboard in the app's settings menu. The keyboard setting unit (11) can control the installation of the third-party keyboard in the user terminal (20) in response to the setting to the enabled state. The installation or activation of the third-party keyboard in the user terminal (20) can be performed automatically or manually with the user's consent.

[0063] The user terminal (20) may be an Android or iOS environment. The keyboard setting unit (11) may allow the user to separately install the third-party keyboard and activate the installed third-party keyboard in either the Android or iOS environment of the user terminal (20), and may enable the use of the service through the activation of the third-party keyboard.

[0064] At this time, the third-party keyboard may be a keyboard developed to minimize cognitive burden by maintaining a UI (User Interface) that is significantly similar to the existing keyboard in an Android environment, and may also be a keyboard developed to aim for a natural user experience by providing a UX (User Experience) similar to the system keyboard in an iOS environment. That is, the third-party keyboard may be a keyboard with a UI / UX similar to the existing default keyboard provided by default in an Android or iOS environment, and through this, the user's sense of unfamiliarity can be reduced (the sense of unfamiliarity can be minimized).

[0065] In addition, in one example of the present invention, the user may use the service by installing the third-party keyboard provided in a form similar to the basic keyboard, but this is not limited thereto. As another example, the device (10) may be implemented (provided) in a form that obtains the user's typing data from the basic keyboard by linking (linking) the device (10) with the basic keyboard already provided within the user terminal (20) without the user having to separately install the third-party keyboard, and thereby enable the user to use the service. In other words, the device (10) may have the aforementioned third-party keyboard applied to it, but it is not limited thereto, and various keyboards such as keyboards previously developed (existing / basic keyboards) or keyboards to be developed in the future may also be applied. That is, regarding the keyboards considered (applied) in the device (10), not only can a third-party keyboard be applied, but if data linkage from the basic keyboard becomes technically possible due to future changes in platform policy, it can be expanded to utilize it.

[0066] The above third-party keyboard may be provided with a plurality of input methods (keyboard layout methods, keyboard types), including, for example, a QWERTY keyboard and a Cheonjiin keyboard, and may be provided (configured) to be usable by a user selectively selecting one of the plurality of input methods. That is, the above third-party keyboard activated on the user terminal (20) may be provided to be activated by the user selecting one of the input methods (one type of keyboard) among the plurality of input methods, including a QWERTY keyboard and a Cheonjiin keyboard. In the above description, the QWERTY keyboard and the Cheonjiin keyboard were only exemplified as being included as types of the plurality of input methods, but this is merely an example to help understand the present invention and is not limited thereto, and various types of keyboards that have been previously disclosed or will be developed in the future (e.g., Naratgul keyboard, Sky keyboard, Vega keyboard, single vowel keyboard, 2-set keyboard, 3-set keyboard, etc.) may be applied. In addition, the keyboards considered (applied) in the device (10) may include not only the aforementioned third-party keyboards, but also various keyboards such as previously developed keyboards (existing / basic keyboards) or keyboards to be developed in the future.

[0067] Once the above-mentioned third-party keyboard is installed and activated within the user terminal (20), the acquisition unit (12) can subsequently acquire the user's digital behavior data from the user terminal (20). In particular, the acquisition unit (12) can acquire the digital behavior data from an application (third-party module, third-party keyboard) installed on the user terminal (20) or from the basic input interface of the operating system (OS) of the user terminal (20).

[0068] Here, the digital behavior data may refer to data (information) regarding various behaviors exhibited by a user who uses a digital device (i.e., a user terminal) such as a user terminal (20). In this case, the word 'behavior' in the present invention may be referred to as 'life reaction,' and thus the digital behavior data may be referred to by terms such as digital life reaction data, daily data, or input data. According to this, the digital behavior data may refer to data regarding various behaviors (reactions) exhibited by a user toward a digital device (user terminal) in their daily life. Such digital behavior data may include the user's typing data, voice data, activity data, etc., and a detailed explanation thereof will be provided later.

[0069] The above digital behavior data can be collected in conjunction with an independent application module, keyboard extension module, accessibility service module, or basic input interface of an operating system operating within a user terminal (20), and can be implemented in a structure that is not dependent on a specific manufacturer's policy.

[0070] The acquisition unit (12) can acquire digital behavior data of the user from the user terminal (20) in real time after the activation of the third-party keyboard, and can store and manage all of the digital behavior data acquired in real time in the database unit (13). Here, the database unit (13) may refer to a memory unit provided within the user terminal (20).

[0071] At this time, the device (10) can control the digital behavior data acquired from the acquisition unit (12) to be processed (stored and analyzed) only locally (i.e., within the user terminal (20)) without external transmission (e.g., transmission to a server).

[0072] Subsequently, the standard setting unit (14) can set (derive) standard behavior characteristic information, which is standard information regarding the user's usual behavior, through the analysis of the digital behavior data obtained from the user terminal (20). Here, the standard behavior characteristic information may be referred to by other terms such as reference point, baseline, etc.

[0073] For example, the standard setting unit (14) may analyze digital behavior data during a pre-set first period among the digital behavior data stored in the database unit (13) to derive user behavior characteristic information corresponding to the first period, and may set the derived user behavior characteristic information corresponding to the first period as the standard behavior characteristic information, which is information regarding the behavior that the user usually exhibits. Accordingly, the standard behavior characteristic information may refer to characteristic (feature) information regarding the user's usual behavior (normal behavior) derived through the analysis of digital behavior data during the first period. Such standard behavior characteristic information may be referred to by other terms such as normal behavior characteristic information. Here, the first period may be set as, for example, 3 months, 6 months, etc., but is not limited thereto, and the setting value for the first period may be set and changed in various ways by the administrator.

[0074] The change detection unit (15) can detect (observe) changes in behavior (i.e., changes in behavioral characteristics) that differ from the user's usual behavior within the additional digital behavioral data obtained after the setting of the standard behavioral characteristic information in the standard setting unit (14) through the analysis of additional digital behavioral data. At this time, the term "to detect changes in the user's behavior" may be expressed differently as "to detect (observe, monitor, detect) changes in behavioral rhythms that differ from the user's usual behavior," "to detect irregularities in daily patterns," "to detect abnormal behavior," etc. Since the device (10) is a technology provided for the purpose of family-centered emotional notification rather than a device or medical device for diagnosing diseases, it can observe (monitor, detect) changes in behavior rather than medically diagnosing (detecting) whether or not changes in behavior (changed behavior) exist.

[0075] For example, the change search unit (15) can analyze digital behavior data during a pre-set second period among the digital behavior data stored in the database unit (13) to derive user behavior characteristic information corresponding to the second period, and then search for changed behavior characteristic information (i.e., changed behavior characteristic information) that is different from the standard behavior characteristic information among the user behavior characteristic information(s) corresponding to the second period.

[0076] Here, the second period may be set as a period following the first period. That is, the second period may refer to a period corresponding to a portion of the time intervals following the first period. For example, the second period may be set shorter than the time interval (time range) of the first period. For instance, if the first period is set to 6 months, the second period may be set to 1 week, 2 weeks, 1 month, etc. However, this is merely an example to aid in understanding the invention and is not limited thereto, and the setting values ​​for the first period and the second period, respectively, may be set and changed in various ways by the administrator.

[0077] For example, when a third-party keyboard is installed in the user terminal (20) on January 1, 2025, the first period may be set to a period of six months (e.g., from January 1, 2025 to June 30, 2025), and the second period may be set to a period of one week after the first period (e.g., from July 1, 2025 to July 7, 2025).

[0078] According to this, the change search unit (15) can identify digital behavior data for a pre-set second period among the digital behavior data stored in the database unit (13) after the setting of the standard behavior characteristic information as the additional digital behavior data. Subsequently, the change search unit (15) can analyze the additional digital behavior data to derive user behavior characteristic information corresponding to the second period, and derive the derived user behavior characteristic information corresponding to the second period as current behavior characteristic information, which is information regarding the behavior currently displayed by the user (current behavior).

[0079] That is, the change detection unit (15) can derive the current behavior characteristic information, which is information regarding the user's current behavior, from the additional digital behavior data through the analysis of the additional digital behavior data. Subsequently, the change detection unit (15) can detect a change in behavior by comparing the current behavior characteristic information derived from the additional digital behavior data with the reference behavior characteristic information. In particular, if the change detection unit (15) analyzes through the comparison that the current behavior characteristic information has a change rate that deviates from a pre-set allowable change range based on the reference behavior characteristic information, it can detect (observe) that a change in the behavior (i.e., change behavior characteristic information) that is different from the user's usual behavior is visible within the additional digital behavior data (or the current behavior characteristic information corresponding to the additional digital behavior data).

[0080] In addition, according to another embodiment of the present invention, the change detection unit (15) may determine the rate of change not only by comparing the user's current behavioral characteristic information with reference behavioral characteristic information, but also by comparing it with group baseline behavioral characteristics calculated from multiple users who have the same age group, gender, or similar characteristics as the user. This method (comparison with group baseline behavioral characteristics) enables the evaluation of relative change, thereby enabling a more stable judgment even when the user's individual baseline is temporarily distorted.

[0081] In particular, the change search unit (15) can identify additional digital behavior data at each second period (e.g., every week) that is pre-set for the time interval after the first period among the digital behavior data stored in the database unit (13) after the setting of the reference behavior characteristic information, and then derive a plurality of current behavior characteristic information corresponding to each of the plurality of additional digital behavior data identified at each second period (i.e., additional digital behavior data at each second period, additional digital behavior data for each second period). For example, the above multiple current behavioral characteristic information may include a first current behavioral characteristic information corresponding to the first second period (e.g., the period from July 1, 2025 to July 7, 2025), a second current behavioral characteristic information corresponding to the second second period (e.g., the period from July 8, 2025 to July 14, 2025), a third current behavioral characteristic information corresponding to the third second period (e.g., the period from July 15, 2025 to July 21, 2025), and the like.

[0082] At this time, the change detection unit (15) can detect a change in behavior within the identified additional digital behavior data by comparing the current behavior characteristic information corresponding to the identified additional digital behavior data with the reference behavior characteristic information whenever additional digital behavior data is identified during the second period. That is, after deriving the plurality of current behavior characteristic information, the change detection unit (15) can detect whether a change in behavior (change behavior characteristic information) is visible within each of the plurality of current behavior characteristic information (i.e., within the additional digital behavior data corresponding to each current behavior characteristic information) by comparing each of the plurality of current behavior characteristic information with the reference behavior characteristic information.

[0083] At this time, the change detection unit (15) can determine (recognize, detect) that a change in behavior is occurring repeatedly in the user for a certain period of time (a certain period of time) or that a different abnormal behavior pattern showing a significant difference from the user's usual behavior pattern is occurring repeatedly, and then control the transmission of notification information through the notification management unit (16).

[0084] That is, the standard setting unit (14) of the device (10) can set the user's standard behavioral characteristic information through the analysis of digital behavioral data during the first period. Subsequently, the change search unit (15) can identify additional digital behavioral data continuously at every second period (e.g., every week) with respect to digital behavioral data acquired in real time after the first period, derive current behavioral characteristic information corresponding to the identified additional digital behavioral data, and search for changes in behavior within the derived current behavioral characteristic information. At this time, let us assume that in the process of repeatedly searching (observing) for changes in behavior at every second period, the change search unit (15) searches (observes) that, for example, among the multiple current behavioral characteristic information, the first current behavioral characteristic information, the third current behavioral characteristic information, and the fourth current behavioral characteristic information show changes in behavior (i.e., changes in behavioral characteristic information), and among the multiple current behavioral characteristic information, the second current behavioral characteristic information and the fifth current behavioral characteristic information show no changes in behavior. In this case, the change detection unit (15) determines that the time when a change in behavior is detected in the fourth current behavior characteristic information is the time when a change in behavior is detected within a number of current behavior characteristic information (e.g., 3) that is pre-set by the manager among the plurality of current behavior characteristic information, and then transmits a control signal to the notification management unit (16) so that notification information can be transmitted through the notification management unit (16).

[0085] That is, the change detection unit (15) does not immediately transmit notification information when a change in behavior is detected in any one of the multiple current behavior characteristic information derived at each second period, but preferably controls the transmission of notification information only when a change in behavior is detected in a number of current behavior characteristic information (e.g., 3) that is pre-set by the manager among the multiple current behavior characteristic information, and then determines that a change in behavior (i.e., abnormal behavior that is a changed behavior different from the usual behavior) is repeatedly occurring in the user. This change detection unit (15) can detect small and repetitive changes in the user's behavior (changes in behavior) in daily life through the analysis of digital behavior data.

[0086] Meanwhile, as previously mentioned, the digital behavior data acquired by the acquisition unit (12) may include the user's typing data, voice data, activity data, etc. Accordingly, the standard behavior characteristic information set by the standard setting unit (14) may include typing-related standard behavior characteristic information, voice-related standard behavior characteristic information, activity-related standard behavior characteristic information, etc.

[0087] <Regarding Typing Data>

[0088] Specifically, the digital behavior data may include typing data of a user input through the keyboard of a user terminal (20). Here, the keyboard may refer to a keyboard corresponding to a third-party keyboard installed on the user terminal (20).

[0089] The standard setting unit (14) can derive a typing pattern (normal, average, typing pattern) regarding at least one of typing speed, rhythm, typing (touch) position, number of typos (typology frequency), and number of deletion inputs through the analysis of the typing data within the digital behavior data, and set it as standard behavioral characteristic information related to typing. That is, the standard setting unit (14) can derive information about the user's typing pattern through the analysis of the typing data, and can set the derived information about the user's typing pattern as standard behavioral characteristic information related to typing (i.e., standard information regarding the user's usual typing behavior). According to this, the derived user's typing pattern may refer to the user's usual typing pattern during a first period.

[0090] Here, the typing data may include raw data of various items (types) as follows. For example, the typing data may include data (data items) such as i) Key Code, which is a unique code of the input key, Press Time, which is the time the key was pressed, Release Time, which is the time the key was released, Dwell Time, which is the time the key was held down, Flight Time, which is the time interval between the time the previous key was released and the time the current key was pressed (i.e., transition time between keys), and Key Sequence, which is the order of all input keys, in relation to data regarding keystroke events. That is, Flight Time may refer to the time taken from releasing one key to pressing the next key. In addition, Dwell Time may be expressed as Hold Times (key sustained input time) as the time a key is held down (i.e., the time taken from pressing and releasing a key).

[0091] In addition, the above typing data may include, in relation to ii) data regarding touch (input location), data such as Touch Coordinates (X, Y), which is information on the exact coordinates where the key was touched; Touch Pressure, which is the pressure intensity of the touch; Touch Area, which is the contact area at the time of touch; and Touch Duration, which is the time the finger stayed on the screen. In addition, the above typing data may include, in relation to iii) data regarding error / correction, data such as the number of Backspace key inputs used for deletion and the number of Delete key usages used to delete characters after the cursor.

[0092] Typing speed may refer to the average time it takes for a user to input a key. The reference setting unit (14) can derive a typing pattern including the average input time per character (or per character) and characteristics (information) regarding the distribution or change (pattern of change of the time) of the time (i.e., the input time per character or per character) by analyzing the typing speed from the typing data.

[0093] Typing rhythm refers to the time pattern of the input interval between each key, and may be information measured by analyzing the time interval between each character when inputting characters. The reference setting unit (14) can derive a typing pattern that includes characteristics regarding the interval pattern of input time between keys (i.e., the pattern of flight time, which is the time interval between keys) and the consistency of input (i.e., consistency of typing regarding whether the typing speed is maintained at a constant level, in other words, consistency of time intervals, etc.) by analyzing the typing rhythm from the typing data.

[0094] The above typing (touch) location may refer to information regarding the coordinates (e.g., (x, y)) that the user actually touched on the keyboard. The reference setting unit (14) can derive a typing pattern including characteristics regarding the error between the actual touch coordinates and the center position of the key (i.e., the error of the touch location relative to the key center), and the distribution density of the touch coordinates by analyzing the typing (touch) location from the above typing data. Here, the error between the actual touch coordinates and the key center may refer to the distance between the center position of the key the user intends to press and the coordinates that were actually touched. In addition, the distribution density of the touch coordinates (density of the touch distribution) may refer to information regarding how concentrated the touch locations are in a specific key or area.

[0095] The above number of typos (typo frequency) may refer to information regarding the number or ratio of typos occurring among all input values ​​entered by the user. The reference setting unit (14) can derive a typing pattern including characteristics regarding the ratio of typos among all inputs and the typo rate in specific keys or patterns by analyzing the number of errors from the typing data.

[0096] The above deletion input count indicates the number of times the user has deleted input, and may be information obtained by measuring the number of times the backspace key is used or the number of characters deleted at once. The reference setting unit (14) can derive a typing pattern including characteristics regarding the frequency of modification after input and the deletion repetition pattern (deletion repetition pattern) by analyzing the deletion input count from the above typing data.

[0097] In summary, the typing-related standard behavioral characteristic information set in the standard setting unit (14) may refer to characteristic information regarding the user's normal typing pattern. This typing-related standard behavioral characteristic information may include information (pattern information) regarding multiple characteristics (features, indicators), such as the average input time per character (or per character), the distribution or change of the time, the interval pattern of input time between keys, the consistency of input, the error between the actual touch coordinates and the center position of the keys, the distribution density of the touch coordinates, the ratio of typos in the total input, the typo rate in specific keys or patterns, the frequency of correction after input, and the deletion repetition pattern.

[0098] Regarding voice data

[0099] Additionally, the digital behavior data may include voice data of the user input through the user terminal (20). In particular, the voice data may refer to voice data of the user obtained, for example, through a call or voice memo on the user terminal (20).

[0100] The standard setting unit (14) can derive a voice pattern (speech characteristic) regarding at least one of the user's speech rhythm (e.g., speech speed, interval, pause interval, etc.), prosodic features (e.g., stress, intonation, tone), and disfluency features (e.g., stuttering, hesitation rate, word repetition, etc.) through the analysis of the voice data within the digital behavior data, and set it as voice-related standard behavior characteristic information. That is, the standard setting unit (14) can derive (or estimate) information regarding the user's voice pattern through the analysis of the voice data, and can set the derived information regarding the user's voice pattern as voice-related standard behavior characteristic information (i.e., standard information regarding the user's usual voice behavior). According to this, the derived user's voice pattern may represent the user's usual voice pattern during a first period.

[0101] The above voice data may include raw data of various items (types) as follows. For example, the above voice data may include i) information related to raw data regarding audio signals, such as PCM data (Pulse-Code Modulation), which is a digitized pure voice signal (Raw waveform); a sampling rate, which is the number of samples per second, typically 16 kHz, 44.1 kHz, etc.; and bit depth, which is generally 16 bits, 24 bits (sound quality precision), etc. Additionally, the above voice data may include ii) information related to acoustic features, such as MFCC (Mel Frequency Cepstral Coefficients) used for speech recognition and classification; a Spectrogram / Mel Spectrogram, which is visualization information according to frequency bands; Pitch (fundamental frequency), which is the speaker's inherent tone; Energy / Loudness, which is the intensity of speech; and Formants, which are resonance frequencies important for distinguishing timbres. Additionally, the above voice data may include iii) information related to metadata, such as recording time and length, and recording device information. In addition, the above voice data may include, in relation to iv) speech-related information (Annotated if labeled), text transcription information at the sentence or word level, information on the speaker's emotional state (e.g., happiness, anger, sadness, etc.), timestamp information at the phoneme and syllable level, and Voice Activity Detection (VAD) information.

[0102] The standard setting unit (14) can derive information (items, characteristics, indicators) regarding speech speed, intervals, and pauses (stops) in relation to the user's speech rhythm through the analysis of the voice data. Here, the pauses may refer to information about the average pause time.

[0103] Here, speech rate may refer to the user's speaking speed derived by measuring the number of syllables or words uttered per second. In other words, speech rate may be information derived based on the total speech length and the number of syllables or words, or it may be derived through consonant-vowel analysis or phoneme-level detection of speech recognition results.

[0104] The above interval (utterance interval) may refer to information regarding the rhythm or flow of utterance derived by analyzing the time intervals between syllables and / or words. The above interval may be information derived by measuring the time intervals between syllables, words, and sentences, and this can be derived by extracting the start and end points of each utterance element within the speech as timestamps.

[0105] The above pause (stop) interval is information derived by detecting a part of speech that is temporarily paused during utterance in voice data, and through the analysis of the pause interval, it is possible to infer the fluency of the speech or the thought process. The above pause interval may be information derived by detecting a silent interval that lasts for a certain period of time or longer within the voice data (e.g., a silent interval of 0.3 seconds or longer), or it may be derived by identifying silence of 300ms or longer through Voice Activity Detection (VAD). Such pause intervals may include information regarding the length and frequency of silence.

[0106] In addition, the standard setting unit (14) can derive information (items) regarding stress, intonation, and tone in relation to the user's prosodic features through the analysis of the voice data.

[0107] Here, the aforementioned stress may refer to information expressed through the intensity and pitch of the voice. In other words, by analyzing the stress occurring in specific syllables or words within the voice data, it is possible to identify which parts the user emphasizes. Specifically, stress may be information analyzed based on whether the relative energy, fundamental frequency (pitch), and duration of a specific syllable rise, or it may be information derived by integrally utilizing these acoustic characteristics.

[0108] The aforementioned intonation refers to pitch variations at the sentence level (e.g., rising intonation, falling intonation). Intonation can be information derived by analyzing the pitch or changes in sound from speech data, and through the analysis of intonation, it is possible to identify information regarding the user's emotional state or manner of expression, such as questions or exclamations. In particular, intonation can be information estimated by analyzing changes in fundamental frequency at the sentence level, or it can be derived by extracting pitch contours at the syllable or word level.

[0109] The aforementioned tone is an item related to the overall tonal mood and emotion, and through the analysis of tone, it is possible to infer the user's emotions and emotional state. In other words, tone can be information analyzed based on the average pitch, energy, and spectral center frequency of the speaker's overall utterance, through which the emotional mood or intention of the speech can be inferred.

[0110] Additionally, the standard setting unit (14) can derive information (items) regarding stuttering, hesitation rate, and word repetition in relation to the user's disfluency features through the analysis of the voice data. Here, stuttering refers to information regarding the number of times stuttering occurs (e.g., 4 times), and word repetition refers to information regarding the word repetition rate (e.g., 2.3%).

[0111] Here, stuttering may refer to information derived by analyzing sections that are repeated or overlapping during speech in voice data (information regarding the rate of stuttering). The hesitation rate may refer to information derived by analyzing sections where unnecessary pauses occur during speech in voice data (information regarding the rate of hesitation). The above word repetition may refer to information derived by analyzing the frequency (ratio) of the same word appearing repeatedly in voice data, and this can help in understanding the user's expression patterns, stress, or the process of reorganizing thoughts.

[0112] In summary, the voice-related standard behavioral characteristic information set in the standard setting unit (14) may refer to characteristic information regarding the user's voice pattern in normal times. This voice-related standard behavioral characteristic information may include information (pattern information) regarding multiple characteristics (features, indicators), such as the user's speech rhythm (e.g., speech speed, interval, pause interval, etc.), prosody characteristics (e.g., stress, intonation, tone, etc.), and disfluency elements (e.g., stuttering, hesitation rate, word repetition, etc.).

[0113] <Regarding Activity Data>

[0114] Additionally, the digital behavior data may include user activity data through the user terminal (20). In this case, the activity data may refer to data (information) regarding the user's activity (use) of the user terminal (20). Such activity data may include usage history data of apps within the user terminal (20), data for detecting inactivity of the user terminal (20), location information (GPS information) of the user terminal (20), etc.

[0115] The standard setting unit (14) can derive at least one of an app usage pattern, an inactive state pattern, and a location pattern through the analysis of the activity data and set it as activity-related standard behavioral characteristic information. Accordingly, each of the app usage pattern, inactive state pattern, and location pattern derived by analyzing the activity data may represent the user's usual app usage pattern, inactive state pattern, and location pattern during the first period.

[0116] At this time, the data derived through the analysis of activity data considered by the device (10) (i.e., user activity-related characteristic information including app usage patterns (e.g., patterns of decreasing app usage), inactivity patterns (e.g., patterns of changes in call volume), location patterns, etc.) may be information (data) indicating signs of emotional distance or social isolation in the user. Considering this, the device (10) seeks information regarding the user's emotional changes or distance through the analysis of activity data and intends to use the sought information to gently share it with family (guardian), etc.

[0117] Here, the app usage pattern may refer to pattern information regarding app usage derived by analyzing usage record data of multiple apps (e.g., phone apps, KakaoTalk, etc.) already installed on the user terminal (20) among activity data. At this time, the usage record data may include, for example, data regarding the usage amount (data usage amount) or usage time of the apps (e.g., total usage time, which is the cumulative time the app was active in the foreground; background usage time, which is the time the app operated in the background without the user's knowledge; last usage time, which is the time the user last used the app). Accordingly, the app usage pattern may include, for example, pattern information of usage time per app and pattern information of data usage per app.

[0118] The above-mentioned inactive state pattern may refer to information regarding a pattern (detection pattern) of a state in which no user activity is detected at the user terminal (20) as the user does not use the user terminal (20) (i.e., inactive state). That is, the inactive state pattern may refer to, for example, pattern information regarding the user's inactive state derived by analyzing inactive detection data among activity data.

[0119] Here, the data for detecting inactivity may include, for example, data regarding the number of missed calls (number of unanswered calls, number of unanswered calls) during a certain period (e.g., a first period) (i.e., data regarding the date and number of missed calls), and record data regarding the time the screen (display) of the user terminal (20) is turned off (or on) due to the user terminal (20) not being used.

[0120] According to this, the above inactive state pattern may include i) pattern information regarding the number of missed calls (frequency) derived by analyzing data regarding the number of missed calls among activity data, ii) pattern information regarding the average time off of the screen of the user terminal (20) (e.g., average time off over a day) derived by analyzing record data regarding the time off of the screen among activity data.

[0121] The above location pattern may refer to pattern information regarding the location of the user terminal (20) derived by analyzing the user's location information (GPS information) obtained through the GPS sensor of the user terminal (20). Such location pattern may include, for example, pattern information regarding the normal movement radius of the user terminal (20) during a first period.

[0122] In summary, the activity-related standard behavioral characteristic information set in the standard setting unit (14) may refer to characteristic information regarding the normal activity (usage) pattern of the user terminal (20). This activity-related standard behavioral characteristic information may include information (pattern information) regarding multiple characteristics (features, indicators), such as usage time per app, data usage per app, the number of missed calls, the average screen off time of the user terminal (20), and location information (especially movement radius).

[0123] Based on this, the change search unit (15) can, after deriving current behavioral characteristic information, compare the derived current behavioral characteristic information with the standard behavioral characteristic information (in particular, the standard behavioral characteristic information related to typing, the standard behavioral characteristic information related to voice, and the standard behavioral characteristic information related to activity, respectively), and if the current behavioral characteristic information is analyzed to have a change rate that deviates from the change allowance range set based on the standard behavioral characteristic information, the change in behavior can be searched (i.e., searched for a change in behavior that is visible).

[0124] In this case, the setting for the aforementioned pre-configured allowable change range may be individually and differently set for each of the multiple characteristics (features, indicators) within the reference behavioral characteristic information, and, for example, may be set in advance by an administrator. Furthermore, the setting for the specific numerical value (threshold) of the allowable change range may be set and changed in various ways by the administrator.

[0125] For example, Mild Cognitive Impairment (MCI) is an intermediate stage between the normal aging process and dementia, referring to a state where there is a slight decline in memory or other cognitive functions (attention, language ability, judgment, etc.), but daily life activities can be performed almost normally.

[0126] In a prior study comparing patients with mild cognitive impairment (MCI) or other early signs of cognitive decline with a healthy control group, it was reported that MCI patients showed a key transition time (Flight Time) that was approximately 117% slower and a key hold time that increased by more than 70% compared to the healthy control group, and that each indicator exhibited high classification performance with AUCs (Area Under the Curve) of 0.997 and 0.976, respectively. In this context, the content regarding the aforementioned prior study mentioned in this invention utilizes the medical characteristics of that study merely for reference purposes, and should not be interpreted in this invention as a diagnostic tool or a basis for medical judgment. In other words, it is clarified that the content regarding the aforementioned prior study should be considered merely as a reference appendix in relation to this invention and is not intended for medical purposes.

[0127] Considering these points, the device (10) may be configured to deliver an emotional notification to a guardian's terminal, etc., when a meaningful change is repeated for a certain period of time (i.e., when additional digital behavior data in which a change in behavior is observed is searched repeatedly three or more times, for example, three times or more) by analyzing the Flight Time and Hold Time in combination among the user's digital behavior data, which is the user's input pattern for a user terminal (20, smartphone), and comparing it with the user's usual reaction characteristics (i.e., comparing current behavior characteristic information related to typing within additional digital behavior data with reference behavior characteristic information related to typing).

[0128] That is, the device (10) can detect changes in behavior by adopting and applying a change recognition algorithm that utilizes the two indicators (Flight Time, Hold Time) described above in combination. At this time, the criteria for determining notification (i.e., the change allowance range that serves as the criterion for detecting changes in behavior) can be configured (set) based on the change rates of the indicators described above, and the specific setting values ​​(threshold values, judgment numerical values) of these criteria can be adjusted in the future by reflecting user feedback and empirical results, and can be set and changed to various values ​​by an administrator. Examples regarding the detection of changes in behavior are as follows.

[0129] < Flight Time Example >

[0130] With this in mind, let us assume that, for example, among the multiple characteristics (features, indicators) included in the above reference behavioral characteristic information, the allowable range of change for the 'characteristic of Flight Time' related to typing data is set to, for example, +30% or less. Here, an increase in the Flight Time value from A to A' can generally be a signal that the typing speed has slowed down.

[0131] At this time, the change detection unit (15) can analyze (determine) that the current Flight Tim (i.e., reference Flight Tim) within the reference behavior characteristic information derived from the reference setting unit (14), for example, is 150ms, and the current Flight Tim (i.e., current Flight Tim) within the user's current behavior characteristic information derived through the analysis of additional digital behavior data is 325ms, and that the current Flight Tim has a change rate (approx. +117%) that deviates from the allowable change range (i.e., +30% or less, or a range of approximately 195ms or less) set based on the reference Flight Tim, and accordingly, can detect that a change in behavior different from the usual behavior is visible within the additional digital behavior data (i.e., current behavior characteristic information). At this time, the change rate of Flight Tim, 117%, may be a value derived by rounding the decimal point in the formula [ (325-150) / 150 x 100 = 116.67% ].

[0132] In other words, if the value of the current Flight Tim falls within the range of 195ms or less, the change detection unit (15) analyzes (determines) that the current behavioral characteristic information has a rate of change that falls within the pre-set allowable change range based on the reference behavioral characteristic information, and in this case, the result can be determined that no change in behavior is visible within the current behavioral characteristic information (i.e., the user is in a normal state). On the other hand, if the value of the current Flight Tim falls within the range of 195ms or more, the change detection unit (15) analyzes (determines) that the current behavioral characteristic information has a rate of change that falls outside the pre-set allowable change range based on the reference behavioral characteristic information, and in this case, the result can be determined that a change in behavior is visible within the current behavioral characteristic information (i.e., the user is in the abnormal state).

[0133] < Hold Time Example >

[0134] For example, let’s assume that among the multiple characteristics (features, indicators) included in the above standard behavioral characteristic information, the allowable range of change for the ‘key hold time (Hold Time) characteristic’ related to typing data is set to, for example, +30% or less. Here, the fact that the value of Hold Time has increased from B to B’ in the past may generally be a signal that the typing speed has slowed down.

[0135] At this time, the change detection unit (15) can analyze (determine) that the current Hold Time has a change rate (approximately +80%) that is outside the change allowable range (i.e., +30% or less, or a range of approximately 260ms or less) set based on the reference Hold Time, and accordingly, can detect that a change in behavior different from the usual behavior is visible within the additional digital behavior data (i.e., current behavior characteristic information).

[0136] < Examples of Typing Speed ​​(Especially, Average Input Time per Character) >

[0137] For example, let’s assume that among the multiple characteristics (features, indicators) included in the above standard behavioral characteristic information, the allowable range of change for the ‘characteristic (pattern) of typing speed (especially, average input time per character)’ related to typing data is set to, for example, +40% or less. Here, the fact that the value of typing speed (average input time per character) has increased from C to C’ in the past may generally be a signal that typing speed has slowed down.

[0138] At this time, the change detection unit (15) can analyze (determine) that the current average input time per character has a change rate (approximately +80%) that is outside the change allowance range (i.e., +40% or less, or a range between approximately 560ms) set based on the standard average input time per character, such that the standard average input time per character is 400ms and the current average input time per character is 720ms. Accordingly, it can detect that a change in behavior different from the usual behavior is visible within the additional digital behavior data (i.e., current behavior characteristic information).

[0139] < Example of stuttering frequency >

[0140] For example, let’s assume that among the multiple characteristics (features, indicators) included in the above standard behavioral characteristic information, the allowable range of change for the ‘characteristic (pattern) of the frequency of stuttering occurrence’ related to voice data is set to, for example, +50% or less. Here, the fact that the value of the frequency of stuttering occurrence has increased from D to D’ may be a signal that one has started to stutter more frequently than usual due to reasons such as a decline in cognitive function.

[0141] At this time, the change detection unit (15) can analyze (determine) that the number of stuttering occurrences in the standard behavioral characteristic information (i.e., the existing average number of stuttering occurrences) is 3 times and the number of stuttering occurrences in the current behavioral characteristic information (i.e., the current number of stuttering occurrences) is 8 times, by analyzing the user's voice data, that the current number of stuttering occurrences has a change rate (approximately +167%) that deviates from the change allowable range (i.e., +50% or less, which is a range of approximately 4.5 times or less, or a range of approximately 5 times or less when viewed based on the rounded value) set based on the existing average number of stuttering occurrences, and accordingly, can detect that a change in behavior different from the usual behavior is visible in the additional digital behavioral data (i.e., the current behavioral characteristic information).

[0142] < Example of missed calls >

[0143] For example, let’s assume that among the multiple characteristics (features, indicators) included in the above standard behavioral characteristic information, the allowable range for change regarding the ‘characteristic (pattern) of the number of missed calls’ related to activity data is set to, for example, +50% or less. Here, the fact that the value of the number of missed calls has increased from the previous E to the current E’ may be a signal that a situation in which the user is unable to answer calls from others through the user terminal (20) more frequently than usual due to reasons such as a decline in cognitive function.

[0144] At this time, the change detection unit (15), for example, can analyze (determine) that the number of missed calls in the standard behavioral characteristic information (i.e., the number of missed calls in the past) is 4 per week and the number of missed calls in the current behavioral characteristic information (i.e., the number of missed calls in the present) is 10 per week, and that the number of missed calls in the present has a change rate (approximately +150%) that exceeds the change allowable range (i.e., +50% or less, or a range of 6 or fewer) set based on the number of missed calls in the past, and accordingly, can detect that there is a change in behavior that is different from the usual behavior within the additional digital behavioral data (i.e., the current behavioral characteristic information).

[0145] < Screen Off Time Example >

[0146] For example, let’s assume that among the multiple characteristics (features, indicators) included in the above standard behavioral characteristic information, the allowable range for change regarding the ‘characteristic (pattern) of the screen off time of the user terminal (20)’ related to activity data is set to, for example, +30% or less. Here, the fact that the value of the screen off time has increased from the previous F to the current F’ may be a signal that the user feels annoyed or uncomfortable about the screen touch or operation itself due to reasons such as a decline in cognitive function, and finds it difficult to look at the screen with concentration, so the time spent looking at the user terminal (20) has decreased compared to usual.

[0147] At this time, the change detection unit (15) can analyze (determine) that the current screen off time (existing / standard screen off time) in the standard behavioral characteristic information is 6 hours per day on average and the current behavioral characteristic information is 10 hours per day on average through the analysis of the user's activity data, and that the current screen off time has a change rate (approximately +66.7%) that deviates from the change allowable range (i.e., +30% or less, which is 1.8 hours, i.e., 1 hour and 48 minutes) set based on the existing screen off time, and accordingly, can detect that a change in behavior different from the usual behavior is visible in the additional digital behavior data (i.e., current behavioral characteristic information).

[0148] As another example, the change detection unit (15) can analyze the user's activity data (i.e., additional activity data) within the additional digital behavior data and, for example, if it is determined (detected) that the screen off time within the additional activity data is longer than a threshold time (e.g., 72 hours) that is preset by the administrator, it can detect that the change in the behavior is visible within the additional digital behavior data.

[0149] < Location Pattern Example >

[0150] The change detection unit (15) analyzes the user's activity data (i.e., additional activity data) within the additional digital behavior data and, for example, determines (detects) that the location pattern (i.e., pattern information regarding the radius of movement) within the additional activity data has not changed for more than a threshold time (e.g., 72 hours) that is pre-set by the manager, that is, in other words, determines that the user's location information (i.e., location information within the additional activity data) obtained through the GPS sensor of the user terminal (20) has not changed for more than the threshold time, and can detect that the change in the behavior is visible within the additional digital behavior data.

[0151] In addition, for example, the change detection unit (15) analyzes the user's activity data and, if it is determined that the pattern regarding the user's movement radius corresponding to the second period (i.e., the pattern regarding the current movement radius) has a rate of change that deviates from the allowable change range (i.e., +30% or less) set based on the pattern regarding the user's movement radius corresponding to the first period (i.e., the pattern regarding the normal movement radius, the normal movement path), it is determined (recognized) that the user has deviated from the normal movement radius (movement path) and is currently living with a movement radius pattern different from the normal one. Accordingly, it can detect that a change in behavior different from the normal behavior is visible within the additional digital behavior data (i.e., current behavior characteristic information). Here, the pattern regarding the movement radius may, for example, refer to the pattern regarding the movement trajectory (movement path, movement path) of the user (user terminal).

[0152] In this way, the change search unit (15) can search for changes in behavior within the current behavior characteristic information by comparing the current behavior characteristic information with the reference behavior characteristic information, taking into account the change allowance range for each characteristic that is individually set differently for each of the multiple characteristics (features, indicators) within the reference behavior characteristic information by the manager.

[0153] Subsequently, the change detection unit (15), as a result of detecting a change in behavior, can determine that if a change in behavior different from the usual behavior is detected within the additional digital behavior data (i.e., current behavior characteristic information), it is an abnormal state in which the user shows signs of mild cognitive impairment (MCI) or other early signs of cognitive decline, and then control (process) the notification management unit (16) to transmit notification information.

[0154] In this example of the present invention, the user in whom a change in behavior is detected (i.e., the user in an abnormal state) is exemplarily described as a person exhibiting signs of MCI or early signs of cognitive decline, but this is merely an example to aid in understanding the present invention and is not limited thereto. As another example, the user in the abnormal state may include not only a user exhibiting signs of MCI (i.e., early signs of cognitive decline), but also a user exhibiting symptoms of dementia, a user exhibiting depression, a user exhibiting postpartum depression, a user exhibiting adolescent anxiety, a user exhibiting attention deficit hyperactivity disorder (ADHD), etc.

[0155] According to this, the standard behavioral characteristic information set in the standard setting unit (14) may include, for example, standard behavioral characteristic information by type of abnormal state. Here, as previously mentioned, the types of abnormal states may include mild cognitive impairment (MCI), dementia, depression, postpartum depression, adolescent anxiety, ADHD, etc. Based on this, the change detection unit (15) can compare and analyze additional digital behavior data and standard behavioral characteristic information based on the standard behavioral characteristic information by type of abnormal state, thereby deriving (identifying) information regarding the type of abnormal state of the user when a change in behavior is detected within the additional digital behavior data (i.e., information on whether the user is a user in an abnormal state) and information regarding the type of abnormal state of the user when a change in behavior is detected, and subsequently, notification information containing the derived information can be transmitted.

[0156] Additionally, when the change detection unit (15) detects a change in behavior that differs from the user's usual behavior, for example, if the current behavior characteristic information derived from additional digital behavior data exceeds a threshold set in correspondence with the reference behavior characteristic information N consecutive times, it can determine this as a change in behavior. Here, N can be a natural number. Additionally, the change detection unit (15) can detect a change in behavior after filtering noise by applying a threshold that can be dynamically changed according to the data collection environment, user-specific characteristics, terminal status, or external noise factors.

[0157] That is, the change detection unit (15) can determine whether there is a real change in behavior while minimizing false positives caused by temporary fluctuations or noise by applying a threshold value that is adaptively adjusted according to the environment, rather than relying on a fixed single reference value.

[0158] In other words, when the change detection unit (15) searches for a change in behavior that is different from the user's usual behavior, for example, by comparing the current behavior characteristic information derived from additional digital behavior data with the reference behavior characteristic information, and determining whether a state exceeding a set threshold is continuously detected, the change in behavior can be detected.

[0159] Here, the threshold value may be dynamically adjusted according to user-specific characteristics, usage environment, data collection conditions, or noise levels, and is not limited to a fixed value.

[0160] The change detection unit (15) can detect gradual or rapid changes in behavioral characteristics by analyzing the time-series trend, slope, speed of change, or acceleration indicator of characteristic values ​​collected over multiple time points. At this time, the change detection unit (15) is not limited to comparing values ​​at a single time point, and can determine the speed and direction of change using the slope, first difference value, second difference value, and moving average-based acceleration indicator of the data at multiple time points.

[0161] That is, when the change detection unit (15) searches for a change in behavior by comparing the current behavior characteristic information derived from the additional digital behavior data with the reference behavior characteristic information, for example, it can search for whether there is a change in behavior by analyzing the trend, slope, or speed of change of characteristic values ​​over time. In searching for a change in behavior, this change detection unit (15) can go beyond a simple numerical comparison at a specific point in time and analyze the slope or acceleration of the behavior data changing over time (i.e., analyze the time-series trend and speed of change of user-specific data) to search for (determine) a change in behavior (anomaly). The 'logic for analyzing slope or acceleration' considered in the present invention can contribute to the advancement of the AI ​​model considered in the present invention.

[0162] For example, the change detection unit (15) can determine that there is a change in behavior when, in the case of typing data, a condition in which the holding time or flight time exceeds a preset threshold value is observed multiple times in succession.

[0163] The change detection unit (15) can be configured to determine a change when, for example, a holding time exceeding about 700ms or a flight time exceeding about 3,000ms is observed two or more times consecutively. However, these examples are merely examples to aid in understanding the present invention, and the threshold value and the number of consecutive observations (e.g., two times) can be variably set according to the application environment or learning model, and can be set and changed in various ways by the manager.

[0164] Additionally, the change detection unit (15) can cross-analyze multiple types of digital behavior data (e.g., typing data, voice data, activity data, walking data, etc.) when detecting a change in behavior. The change detection unit (15) can detect a change in behavior by cross-analyzing multiple indicators, such as voice, activity, and walking data, in addition to typing.

[0165] The notification management unit (16) may transmit (send) notification information indicating that there is a change in the user's behavior when the change in the behavior is detected. Here, the notification information may refer to notification information intended to induce emotional interest from family (or guardian), etc., rather than inducing medical measures. Therefore, the notification information may be referred to by other terms such as emotional notification information, emotional reference notification information, etc. At this time, the notification management unit (16) may transmit the notification information using at least one of text messages (SMS), push notifications (pop-up windows), and messenger apps (e.g., KakaoTalk, etc.). Such notification information may be referred to by other terms such as emotional notification information.

[0166] The notification management unit (16) can generate notification information when a change in behavior is detected (observed) and then transmit the generated notification information to each of the terminals of the notification recipients who are pre-registered in relation to the user. The notification information may be, for example, information such as, "It appears that early signs of cognitive decline are appearing in the user, so attention and verification are required."

[0167] Here, the notification recipients may include, for example, first-type notification recipients registered by a user who has pre-registered the device (10) through a user terminal (20), and second-type notification recipients registered by an administrator who has pre-registered the device (10) through an administrator terminal (not shown). At this time, the first-type notification recipients may include, for example, a user, at least one guardian (a guardian who is a family member of the user), and a caregiver who manages and is responsible for the user's care. Additionally, the second-type notification recipients may include, for example, a person in charge of a local government or welfare official managing the user (i.e., a person in charge of a public account), a control center, etc.

[0168] For example, when a change in behavior is detected (observed), the notification management unit (16) first transmits the generated notification information to the terminals of the first type of notification recipients registered by the user, and then ii) if a first time (e.g., 24 hours) has elapsed after the first time has elapsed, or if it is detected that all of the first type of notification recipients have not checked the notification information within a second time (e.g., 6 hours) after the notification information has been transmitted first, then it can secondarily transmit it to the terminals of the second type of notification recipients registered by the manager.

[0169] Additionally, the notification management unit (16) can transmit the generated notification information to the terminals of the first type of notification recipients and the terminals of the second type of notification recipients simultaneously or sequentially when a change in behavior is detected (observed). That is, the notification management unit (16) can control, for example, a user or administrator to pre-register multiple notification recipients in the device (10), including not only the user's family (guardian) but also caregivers, local government officials, and welfare officials, and can transmit the notification information to each notification recipient simultaneously or sequentially.

[0170] Additionally, the notification management unit (16) can transmit the generated notification information to the terminals of the second type of notification recipients, for example, when the user has not pre-set (registered) the first type of notification recipients. According to this, the notification management unit (16) can transmit the notification information to the terminals of the second type of notification recipients who have been pre-registered by the manager in preparation for a situation where the user has not registered the guardian's contact information (mobile phone number) in the device (10), or where it is determined that the first type of notification recipients cannot be contacted because they have not all checked the notification information within a pre-set second time (for example, 6 hours) even though they have registered it.

[0171] The notification management unit (16) can control the transmission order or transmission method (i.e., simultaneous or sequential method) or transmission target of the notification information for the notification recipients so that the user can pre-set (directly set) the setting information in the device (10). That is, the notification management unit (16) can provide the user with a setting function for the setting information and, based on this, can control the transmission of the notification information based on the setting information pre-set by the user.

[0172] The notification management unit (16) can transmit notification information to the terminals of the notification recipients when the change in behavior is detected (observed) as described above. In addition, the notification management unit (16) can transmit additional guidance notification information to the terminal of at least one of the notification recipients (multiple notification recipients) to induce regular checks or attention to the user when the change detection unit (15) detects a change in behavior within additional digital behavior data, and when the change in behavior is not detected (observed) during a monitoring period (e.g., 8 weeks) pre-set by the manager after the standard behavior characteristic information is set (i.e., when no significant difference in reaction (behavior) from the current behavior compared to the usual behavior is seen (observed) during a predetermined period). For example, the guidance notification information may include information such as, "Please take an interest in whether anything is wrong with the user."

[0173] The control unit (17) can control the operation of each part (i.e., keyboard setting unit, acquisition unit, database unit, standard setting unit, change search unit, and notification management unit) within the device (10). Additionally, the control unit (17) can control the operation of the user terminal (20).

[0174] Additionally, the control unit (17) can control the main app of the device (10) installed on the user terminal (20) (i.e., the app provided by the device) to transmit a normal operation signal (Ping) to the server once a day (1 time), for example. In this case, the server (not shown) can transmit the guidance notification information, which is a notification created for the purpose of emotional attention arousal, to the terminal of a guardian (family) who is registered in advance by the user, if the normal operation signal is not received from the user terminal (20) (especially from the main app within the user terminal) for more than a threshold time (e.g., 72 hours) that is pre-set by the administrator.

[0175] For example, the control unit (17) may be configured to automatically send a normal operation signal to the server at a time pre-set by the administrator once a day (e.g., 1:00 AM) when the app is running normally, a network connection is available, and the app is designed to operate regardless of whether it is in the background or foreground. At this time, the control unit (17) may determine that the app is not operating normally when, for example, the battery of the user terminal (20) is discharged and the power of the user terminal (20) is turned off, the app is forcibly terminated or deleted from the user terminal (20), or the user terminal (20) is not connected to a network (Internet), and in this case, the normal operation signal cannot be sent to the server.

[0176] Accordingly, if the server (not shown) does not receive the normal operation signal from the user terminal (20) for a period longer than the preset threshold time, it determines (infers) that the user terminal (20) is turned off or that the user's behavior (use, activity, response) regarding the user terminal (20) is in a state of long-term non-action (non-use, non-activity, non-response), and can subsequently transmit the guidance notification information to the guardian's terminal so that the guardian may take an interest in the user. According to this, the server (not shown) can provide a function capable of detecting the user's non-response (abnormal signs) based on the normal operation signal (Ping). In other words, the server (not shown) can detect (detect) a state in which the normal operation signal from the user terminal (20) has not arrived for a certain period of time (72 hours) or longer, and in this case, determine that the user has not used the user terminal (20) for a long time or that the user is absent, and then control the transmission of the guidance notification information, which is a notification to induce attention so that the guardian can take an interest in the user, to the guardian's terminal.

[0177] At this time, the Ping-based non-response (abnormal sign) detection logic considered by the server can be applied (implemented) as a bypass structure to avoid background restrictions when, for example, the user terminal (20) is in an iOS environment, and can be applied (implemented) as a local-based detection structure combining location, screen events, etc. when the user terminal (20) is in an Android environment. In other words, the device (10) controls the Ping-based non-response detection to be performed in a way that bypasses these restrictions, i) when the user terminal (20) is in an iOS environment, since there are restrictions preventing apps from operating well in the background in the iOS environment, and ii) when the user terminal (20) is in an Android environment, since the restrictions are less severe than in the iOS environment, it can control the non-response detection to be performed locally without Ping by using events such as location, screen touch, etc.

[0178] According to this, the device (10) can detect behavioral changes based on digital behavior (digital behavior data) of a user terminal (20) by combining local-based processing (analysis processing) and server-based detection (detection of non-response from the server).

[0179] As described above, the device (10) can detect usage trends different from usual (changes in behavior, which are usage trends of the user's device different from usual) based on data such as typing speed, location movement (location pattern), phone response (e.g., number of missed calls or screen off time), and app usability (i.e., app usage pattern), which are daily data (digital behavior data) naturally generated during the process of the user using the user terminal (20, smartphone). Through this, the device (10) can transmit notification information for emotional reference to family members or guardians, etc., if the user's change (i.e., change in behavior, change in behavior) is repeated for a certain period of time or shows a significant difference from the usual pattern, even if the user's change (i.e., change in behavior) is not intended for medical judgment or disease diagnosis in providing this service, but its main goal is merely to help the user's daily changes be shared earlier and more naturally within a warm care system centered on family (user's guardian, etc.).

[0180] The device (10) can detect early signs of mild cognitive impairment (MCI) or other cognitive decline in a non-medical manner and transmit the detected information to a guardian or other person in the user's vicinity as an emotional notification. The device (10) may also be referred to by other terms such as a smartphone-based daily change response transmission device, a smartphone-based emotional notification device for detecting (observing) changes in behavior regarding the user terminal (20, smartphone) due to the user's cognitive decline, etc.

[0181] The device (10) is intended not to diagnose a user's disease / disease (e.g., dementia or mild cognitive impairment) or provide a medical interpretation, but rather to detect and convey the trend of change in the user's life responsiveness to the user terminal (20). That is, the device (10) is not a medical diagnostic tool, nor is it a device that provides a disease classification function. The device (10) may be a device that provides notification information for the purpose of arousing emotional interest to a guardian, etc., through trend detection based on the user's standard change rate (i.e., by exploring changes in behavior based on the rate of change of behavior based on the user's standard behavioral characteristic information). The device (10) may be a device for emotional notification for the purpose of protection (prevention, observation) regarding the occurrence of a user's disease / disease, rather than diagnosis.

[0182] The device (10) can detect changes in the user's behavioral response based on typing data, detect changes in repetition and rhythm based on voice data, and detect non-response (e.g., a sharp decrease in app usage, an increase in the number of unanswered calls, location deviation, etc.) based on activity data.

[0183] The device (10) can detect changes in the user's behavior based on the user's typing data. In this regard, the device (10) can compare and analyze the user's digital input patterns (typing-related input patterns), such as typing speed, rhythm, and touch location, over a certain period. If a tendency different from the user's unique standard (i.e., typing-related standard behavioral characteristic information) (i.e., a change in behavior different from usual behavior) is repeated, the device (10) can transmit information (notification information) indicating that the different tendency is repeated to a guardian or designated account (e.g., second-type notification recipients) in the form of a reference guidance message. According to this, the device (10) can be said to have a structure that accepts (judges) the user's typing pattern (e.g., pattern of typing speed, pattern of rhythm, etc.) as a signal (a signal indicating that there is a possibility of abnormal signs such as cognitive decline in the user) rather than a simple typo, and transmits the relevant information to a guardian, etc., when the user's typing pattern (e.g., pattern of typing speed, pattern of rhythm, etc.) repeats a pattern different from the previous pattern.

[0184] The device (10) can analyze patterns such as the user's typing speed, rhythm (typing input rhythm), and touch location, and if a trend that is repeatedly different from the user standard set based on previous input data (i.e., typing-related standard behavioral characteristic information set through the analysis of typing data during the first period) is observed, the device (10) can transmit the information to a guardian. In other words, if the current typing-related current behavioral characteristic information derived by analyzing the user's typing data during the second period is detected to have a rate of change that exceeds the allowable range of change set based on the user's typing-related standard behavioral characteristic information corresponding to the first period, the device (10) can detect that a change in behavior different from the user's usual behavior is being observed and transmit the detected information to a guardian, etc.

[0185] FIG. 2 is a diagram schematically illustrating the processing of typing data in a digital behavior-based behavior change detection device (10) according to one embodiment of the present invention.

[0186] Referring to FIG. 2, the device (10) can acquire (collect) input events (i.e., typing data generated by the user operating the third-party keyboard) through a third-party keyboard installed on a user terminal (20). At this time, the acquired typing data may include data regarding keystrokes, backspaces, hold times, flight time, etc. The device (100) can analyze typing patterns including typing speed, typo occurrences, and patterns of deletion by analyzing the acquired typing data, and if the typing pattern analyzed this time (current typing pattern) shows a change rate of a certain level or more compared to the previous typing pattern a certain number of times or more, notification information can be provided to the guardian.

[0187] In the device (10), the detection of whether a change in behavior is observed within the current behavioral characteristic information (i.e., additional digital behavioral data) (i.e., detection of whether the user is in an abnormal state) can be performed based on a relative change rate for each user that is set relatively for each user, rather than a fixed reference value. In other words, the change allowance range (i.e., a change allowance range set based on reference behavioral characteristic information) that serves as the criterion for detecting a change in behavior in the device (10) is set by an administrator, but can be set differently for each user rather than as a fixed value.

[0188] In addition, the device (10) can perform automatic learning based on user standards, and in this regard, it can set a standard value (standard behavior characteristic information) for each user based on initial input data (i.e., digital behavior data during the first period), and then improve the accuracy of detection by learning and correcting the standard value through continuous learning using subsequent input data (i.e., additional digital behavior data).

[0189] At this time, the user typing data considered in the device (10) may be data collected through a third-party keyboard installed on the user terminal (20) with the user's consent in the Android and iOS operating systems, as previously mentioned. However, it is not limited to this, and the user typing data considered in the device (10) may also be data collected through a basic keyboard provided by default in the user terminal (20) or an input interface provided on an existing operating system.

[0190] In addition, the device (10) can detect changes in the user's behavior based on the user's voice data. In other words, the device (10) can detect changes in the user's daily response (changes in behavior) based on changes in the user's voice pattern by analyzing the user's voice data. In this regard, the device (10) can analyze the user's voice patterns, such as speech speed, rhythm, and repetition, and if a repetitive difference from the user-specific standard is identified (detected), the information can be transmitted to a guardian, etc. The device (10) can analyze (detect) changes in the user's voice pattern regarding elements such as speech speed, intervals between syllables, stuttering, word repetition, and breakdown of speech rhythm through the analysis of acquired voice data. The device (10) can analyze the user's voice data acquired based on calls or voice memos from the user terminal (20), and if changes in speech speed, intervals between words, and non-verbal elements of the voice (stuttering, hesitation rate, etc.) are repeated beyond a certain level, the trend can be transmitted to the guardian. This device (10) is not intended for diagnostic purposes, but rather to transmit information (detection information) about the flow of changes in the user's speech patterns, etc., to a guardian, etc.

[0191] That is, the device (10) can detect that a change in behavior is observed in the current user and transmit that information to a guardian, etc., when the current voice-related current behavioral characteristic information derived by analyzing the user's voice data during the second period is detected to have a rate of change that exceeds the allowable range of change set based on the user's voice-related standard behavioral characteristic information corresponding to the first period.

[0192] At this time, the device (10) may be provided in the form of an SDK (Software Development Kit) to enable an add-on to a call app, a message app, etc., that is already installed within the user terminal (20) in order to acquire the user's voice data from the user terminal (20). The app provided by the device (10) may be provided, for example, in the form of an SDK, to be added to a call app, etc., within the user terminal (20) and configured to operate when the app is activated. In other words, the device (10) may be provided in the form of an SDK (Software Development Kit) so that it can be linked to a call app or a message app of the user terminal (20). At this time, the SDK provided by the device (10) may be configured (provided) to be integrated into the app in the form of an add-on, and designed (provided) so that voice data can be automatically collected (acquired) when the app is activated. Here, an SDK refers to a set of tools (software development kit) that helps to easily add new functions or link with a specific platform or app. Additionally, an add-on refers to an additional element that is attached to software to extend its functionality. Furthermore, the device (10) can be provided in the form of an SDK or PDK that can be linked to third-party services, in addition to the form of an app. That is, the device (10) has the potential to be expanded into the form of a Software Development Kit (SDK) or a Process Design Kit (PDK).

[0193] The device (10) can acquire voice data by linking with an external app (call app or message app) in the form of an SDK, and can perform analysis and processing operations on the acquired voice data under the principles of local analysis and non-identification processing. The device (10) can be provided in the form of SDK linkage with other call apps for acquiring voice data, and can ensure a level of personal information protection by ensuring that analysis is performed only locally when analyzing voice data. That is, to protect the user's privacy, the device (10) can control that, among the digital behavior data acquired by the acquisition unit (12), the user's voice data, in particular, must be analyzed and processed only locally, and can control that the acquired voice data (recording files, etc.) is not transmitted externally (e.g., a server). The device (10) can control that voice data is processed non-identified locally and can control that the recording is not transmitted.

[0194] The device (10) can ensure that all digital behavioral data (i.e., digital behavioral data including typing data, voice data, and behavioral data) obtained from a user terminal (20) is processed locally and can be processed so that it is detected in a de-identified state without transmission to or storage to a server. In particular, the device (10) can process voice data so that it is not recorded or transmitted externally. The device (10) can be said to have a design structure that protects the user's privacy (personal information), thereby preventing the possibility of classification as a medical device and ensuring legal stability.

[0195] Additionally, the voice data acquisition (collection) function of the device (10) can be designed (provided) to operate on both Android and iOS. The device (10) can automatically acquire the user's voice data when the user is using the app or is in a call by automatically detecting when the user is using the call app or message app within the user terminal (20) or is in a call.

[0196] In addition, the device (10) can detect changes in the user's behavior based on the user's activity data. In this regard, the device (10) can detect trends in changes in the user's lifestyle reactions (i.e., detect changes in behavior) by analyzing user activity data such as the number of missed calls, location patterns (location deviations), and app usage patterns (changes in app usage patterns), and can ensure that the information is delivered to a guardian. That is, the device (10) has a structure that can comprehensively determine states such as non-use of the device (user terminal) for a certain period of time, no change in location, and non-connection to apps by analyzing activity data in addition to typing or voice data, and provides notification information to a guardian or public account based on this.

[0197] The device (10) can detect that a change in behavior is observed in the current user and transmit that information to a guardian, etc., when the current behavioral characteristic information related to the current activity, derived by analyzing the user's activity data during the second period, is detected to have a rate of change that exceeds the allowable range of change set based on the standard behavioral characteristic information related to the user's activity corresponding to the first period. At this time, the device (10) can detect (analyze) information regarding the current behavioral characteristic information related to the current activity, such as whether the number of missed calls by the user has increased rapidly within a certain period, whether the usage time of a specific app has decreased rapidly, or whether there has been no change in the location information of the user terminal for more than a preset threshold time (e.g., 72 hours), by analyzing the activity data, for example, in comparison with the standard behavioral characteristic information related to the user's activity, and if such detection occurs, control can be made so that notification information is transmitted to a guardian, etc.

[0198] The device (10) can analyze (determine) changes in the number of missed calls, deviation from the location radius (i.e., deviation from the movement radius), and app usage patterns through the analysis of activity data. Through this, it can detect information such as whether there is a difference in the current movement path (movement radius) compared to normal times (i.e., whether it has deviated from a certain radius), whether the number of missed calls has increased rapidly (surge) within a certain period, and whether the app usage pattern has decreased rapidly. When such detection occurs, it can detect that there is a change in the current user's behavior compared to normal times in the past and provide notification information thereafter.

[0199] Typically, people who have patients with dementia or mild cognitive impairment around them often express regret, saying things like, "I should have known back then..." Accordingly, the service provided by the device (10) was developed to create more opportunities to take an interest at least once, even though it is not possible to perfectly identify that moment.

[0200] The device (10) aims to induce a single moment of attention and create an opportunity for a cautious connection by exploring changes in the flow of digital behavior that naturally occur during smartphone use by a user with potential for cognitive change, and by transmitting this to a guardian or public account when a daily pattern different from usual is repeated for a certain period.

[0201] The device (10) detects the flow of changes in the user's daily life (behavior) response to the user terminal (20) based on digital behavior data of the user (i.e., data related to information flow such as the user's typing speed, input rhythm, and location movement) obtained through the user terminal (20) without the need for separate sensors or equipment, and provides a technical structure that can listen to small signals (signals of behavioral changes) in daily life without the purpose of diagnosis. In other words, conventional systems for detecting the user's behavioral response were mainly composed of hospital devices or wearables, which had the disadvantage of requiring separate equipment. In contrast, the device (10) can effectively detect the flow of changes in the user's daily pattern using the app (or third-party keyboard) installed on the user terminal (i.e., smartphone) possessed by the user without the need for such separate equipment, and thus can be said to have the advantages of high accessibility and a low cost structure compared to conventional technologies.

[0202] The device (10) is designed so that each function within the device (10) operates (for example, so that digital behavior data is acquired) through the normal use of a smartphone (user terminal) without any separate operation. A third-party keyboard installed on the user terminal (20) can be activated automatically or manually with the user's consent. The device (10) is not a medical device but a non-invasive tool for exploring daily information flows, and has the advantage of being able to be distributed to more users thanks to its mobile-centric accessibility and low-cost structure.

[0203] In addition, the device (10) has a notification structure that, when transmitting notification information, transmits the notification information to the guardian's terminal if the guardian's information (contact) is pre-registered, thereby enabling family-centered notifications. Conversely, when the guardian's information is not pre-registered, it has a notification structure that transmits the notification information to the local government or public account through linkage with the local government or public account. This device (10) can minimize gaps in the social safety net through the aforementioned notification structure. The device (10) can be expanded into a complex care model that encompasses B2C (Business to Consumer), B2B2C (Business to Business to Consumer), and B2G2C (Business to Government to Consumer) structures. Here, B2C refers to services between a company and a consumer. B2B2C refers to a structure where one company provides products or services to another company, and that company provides them to the end consumer. B2G2C refers to a structure where a company provides technology or solutions to the government, and the government utilizes those solutions to provide services to the general public.

[0204] In addition, the device (10) can improve both technical precision and the depth of care for the user through various advancements such as user-specific reference point (reference behavioral characteristic information) learning, guardian app advancement, participatory reward system, and voice-based functions.

[0205] In particular, considering the situation where the guardian (family, etc.) is slow to recognize the occurrence of phenomena such as cognitive decline in the user, the device (10) can create an opportunity for the guardian to take an interest in and converse with the user through simple pattern transmission (for example, transmission of notification information that a pattern such as typing speed has changed), and can provide an environment that leads to a flow in which interaction between the user and the family increases and reactions within the care system are promoted.

[0206] Changes in a user's daily routine are not limited to the individual but can lead to emotional, temporal, and social burdens on surrounding family and community. In particular, recognizing and responding to the initial flow where the severance of relationships begins a little earlier may be a more critical point than memory loss. Accordingly, the device (10) is significant in that it captures the beginning of such relational distance and creates an opportunity for attention regarding it. The device (10) may be a device developed for the purpose of emotional warning to help guardians (family, etc.) take an interest in the user more quickly if changes in the daily routine persist for a certain period, rather than being a medical diagnostic or preventive tool. In other words, the device (10) is not a service provided for the purpose of medical diagnosis or disease classification, but is intended to provide a catalyst for emotional attention to guardians.

[0207] The present invention can be said to differ from the official tests or examination methods used in existing long-term care insurance screenings. Since the device (10) operates based on the flow of naturally using a smartphone (user terminal) in daily life without the user having the awareness that they are being evaluated, it can minimize cognitive distortion known as the Hawthorne Effect. The purpose of the device (10) is not to induce or test specific behaviors, but to quietly detect and transmit the flow of change within the user's daily reactions. That is, the device (10) can acquire digital behavior data, including the user's typing data, through a third-party keyboard installed on the user terminal (20), and analyze this data to explore changes in behavior. In this way, the user is prevented from having the awareness that they are being evaluated when acquiring digital behavior data, and through the analysis of the digital behavior data that occurs naturally (acquired), changes in behavior that have changed significantly compared to normal can be effectively detected (observed, captured).

[0208] The device (10) quietly monitors the small flow of daily life (flow of behavior) while the user naturally uses a smartphone (user terminal) without being aware of it, and when the change is repeated (i.e., when the change in behavior is repeated more than a certain number of times), it transmits notification information to the recipients of the notification informing the user that a change in behavior different from their usual behavior has been observed or detected, thereby enabling the starting point of care for the user to be brought forward a little more. This device (10) can play a role in helping a small safety net to operate at a slightly earlier point (time) than the existing social safety net.

[0209] That is, since the device (10) can detect changes in behavior through the analysis of typing data obtained from a third-party keyboard installed on a user terminal (20), it is possible to continuously (real-time) observe (search) whether changes in behavior are visible naturally without the user being aware (perceived) that they are evaluating (detecting, observing, searching) changes in their own behavior, and it is possible to avoid the inconvenience of the user having to fill out a separate survey when searching for changes in behavior.

[0210] In addition, in South Korea, where smartphone penetration is high, this device (10), which encompasses both Android and iOS environments, can establish itself as a technology of ‘small warmth’ that reduces blind spots together between the private and public sectors.

[0211] In addition, the device (10) can minimize psychological stress or defensive reactions that may occur during the use of the user terminal (20) by delivering notifications regarding digital response-based detection results (i.e., results of exploring behavioral changes based on digital behavioral data) in a manner that considers the emotional stability of the user and guardian, rather than in the form of danger warnings (i.e., delivered as notification information for emotional reference). This device (10) is not a system that diagnoses diseases or performs mandatory monitoring like a medical device, but is entirely non-medical and can be used as a cautious connection device that naturally "awakens attention" within the flow of the user's daily life. The device (10) is intended for relationship restoration and attention inducement rather than surveillance or control, and can be described as a lifestyle notification technology designed on a philosophy fundamentally different from surveillance systems like "Big Brother." In other words, the device (10) can be described as a non-medical device-based structure that explores trends of repetitive lifestyle response changes rather than for diagnostic purposes, and can be used as a care judgment aid tool rather than for diagnostic purposes.

[0212] The device (10) has the potential to be applied to the field of academic research that analyzes the correlation between behavioral patterns and cognitive states based on de-identified data, provided that user consent and collaboration with research institutions are achieved in the future.

[0213] In addition, the device (10) may initially be implemented as a Minimum Viable Product (MVP), and subsequently, the following advanced plan may be established and applied to the device (10) to improve usability and scalability. That is, the device (10) may be advanced after the MVP implementation.

[0214] <Application Scenario of the Present Invention>

[0215] The device (10) can be applied in, for example, the following environment, and can be configured to be expandable to the following scenarios in the implementation stages according to the combination of components.

[0216] 1) Basic (MVP) Implementation Phase

[0217] The device (10) may be configured based on a user app (a user-exclusive app available to the user) and a third-party keyboard during the MVP implementation phase. In this regard, the device (10) may allow the user to install the app (user-exclusive app) on the user terminal (20) and subsequently set the activation status of the third-party keyboard included in the app to an activated state, and may acquire (collect) the user's typing data through the activated third-party keyboard. In addition, the device (10) may acquire (collect) digital behavior data including typing data from the user terminal (20). The device (10) can acquire typing data in real time based on full access rights to a third-party keyboard, and can detect (search) changes such as typing speed, rhythm (input rhythm), and frequency of typos through analysis of the acquired typing data, and if a tendency different from usual is repeated beyond a certain level (i.e., if an abnormal state in which the current behavioral characteristic information related to typing has a rate of change that exceeds the pre-set allowable range of change based on the standard behavioral characteristic information related to typing is repeated multiple times (e.g., 3 times or more), it can generate notification information to inform of this and provide it to a guardian, etc.

[0218] In addition, the device (10) can detect changes (explore changes in behavior) through comparison with a reference point (i.e., reference behavioral characteristic information). In this regard, the device (10) can deliver notification information as a reference alert to a guardian when a change in input pattern exceeding a certain level is detected based on a user-specific reference point (baseline). The device (10) can be utilized (applied) in both iOS and Android environments and can be configured to operate based on a third-party keyboard in both iOS and Android environments. In particular, the device (10) can be configured to be designed with a structure suitable for iOS environments, as iOS has relatively stronger security and background restrictions compared to Android when the user terminal (20) is in an iOS environment.

[0219] In addition, the device (10) can determine an abnormal state (a state in which a change in behavior is observed) based on the location information of the user terminal (20). In this case, in the case of an Android device, the app can detect the location and screen turning on of the user terminal (20) even in the background. Taking this into consideration, the device (10) can, for example, when the user terminal (20) is in an Android environment, if the screen turning off time in the user's activity data is detected to be longer than a preset threshold time (e.g., 72 hours) or if the location pattern in the activity data is detected to be unchanged for longer than a preset threshold time, recognize (determine) that the user's inactive state regarding the user terminal (20) has continued for longer than the said threshold time (72 hours), and subsequently control the sending of the notification information to a guardian, etc., as a reference notification regarding the difference in daily response.

[0220] In addition, when the device (10) transmits the notification information or the guidance notification information to the notification recipients during the MVP implementation phase, it may transmit the notification information or guidance notification information using at least one of a text message (SMS) notification, a push (popup window) notification, or a messenger app (e.g., KakaoTalk, etc.) notification, without, for example, a notification recipient-only app (e.g., a guardian-only app, a local government official-only app, etc.).

[0221] 2) Advanced stage

[0222] In the advanced stage, the device (10) may create (develop) a guardian-only app separate from the user app (i.e., user-only app) and provide it to the guardian. The guardian may install the app (i.e., guardian-only app) using a terminal they possess (guardian terminal). The guardian-only app may be linked with the device (10) and the user terminal (20), and based on this, it may be provided with a function to periodically check the status of the user (i.e., a function to periodically check the user's status) and a function to receive notification information related to the user. Additionally, the guardian-only app may be provided with a function to allow the guardian to register notification recipients who receive notification information related to the user (i.e., a function to register multiple notification recipients), a function to check the user's abnormal history, and a function to link with a hospital.

[0223] That is, the device (10) can have its functions expanded by introducing a dedicated app for guardians. The device (10) can be configured such that the dedicated app for guardians has a dashboard function to check user status (e.g., recent activity flow and notification history), a function to register multiple users (e.g., multiple notification recipients such as family members and caregivers), a function to set permissions and share information among guardians, and a function to connect (communicate) with a hospital (nursing facility, etc.) or local government by adding a linkage module with a hospital or local government. This device (10) can be expanded beyond a B2C-based model to a B2B2C model targeting nursing facilities.

[0224] The device (10) can analyze user voice data obtained from a user terminal (20) based on a call or voice memo to analyze trends in voice pattern changes such as speech speed, interval, and hesitation rate, and can transmit relevant information to a guardian when the detection of changes in voice patterns is repeated more than a certain number of times. In addition, the device (10) can notify the guardian of information regarding the flow of behavioral changes related to specific diseases (e.g., MCI, dementia, depression, postpartum depression, etc.) through the analysis of trends in voice pattern changes. The voice pattern analysis-based behavioral change detection function applied to the device (10) may or may not be included in, for example, the MVP and public distribution version, and may be developed and applied in the future through medical consultation or corporate collaboration structures.

[0225] In addition, the device (10) can provide a Ping-based non-response detection function (logic) as previously mentioned.

[0226] In addition, the device (10) can be configured to provide the aforementioned notification information or the aforementioned guidance notification information to the second type of notification recipients by linking with a social safety net when there are no first type of notification recipients (i.e., guardians or caregivers, etc.) registered by the user or guardian. For example, when the user using the service has not registered guardian information (i.e., in the case of a user without guardian information), if digital behavior data is not acquired from the user terminal (20) for a certain period of time (e.g., 72 hours) or longer, and the user terminal (20) detects that the user is in a state of non-response (non-use) for a long period, the device (10) can be configured to provide the aforementioned notification information or the aforementioned guidance notification information to the terminals of second type of notification recipients, including a control center or a welfare official in charge. Compared to conventional high-cost installation systems that include motion detectors or door sensors for detecting abnormal user conditions, this device (10) can be utilized as a significantly low-cost, non-installation alternative.

[0227] That is, the device (10) can be configured to be linked to a public account-based social safety net. Through this, the device (10) can be configured to form a basic social safety net delivery structure by linking accounts of local governments or welfare officials for users without guardians. This device (10) has the structure of a B2G2C model and can be designed and utilized as a non-installation, low-cost alternative that can replace existing high-cost installation equipment.

[0228] The device (10) has an optional linkage structure configured to enable the transmission of supplementary information to registered public accounts (i.e., second-type notification recipients) according to the settings or policies of the user or administrator when a change in trend is continuously confirmed (more than a certain number of times) from the user. The device (10) has a supplementary function to prepare for cases where the user has no guardian or cannot be contacted, and can be provided to have an optional social safety net expansion structure utilizing existing infrastructure.

[0229] That is, the device (10) has a structure that registers multiple guardians and is linked to the public. The device (10) can be configured to send notification information to a local government or welfare official account in case the user has not registered a guardian or cannot be contacted. In addition, the device (10) is configured to have a setting function that allows multiple notification recipients (family, caregivers, etc.) to be registered and to send notification information to these multiple recipients simultaneously or sequentially. Such a device (10) can contribute to preventing lonely deaths of users and strengthening the social safety net.

[0230] In addition, the device (10) can be configured to enhance the search structure for multiple behavioral pattern changes. In this regard, for example, i) in the initial MVP implementation stage, it can be configured to search for changes in user behavior (search for changes in behavioral tendencies) based on a single behavioral characteristic such as typing speed, typing position, number of missed calls, etc., and ii) in the future, through advancement, it can be configured to search for changes in user behavior based on a complex search logic that combines various elements (behavioral characteristic elements) such as typing rhythm (input rhythm), repetition, position movement pattern, screen lighting / usage time, etc. At this time, each indicator used to search for changes in user behavior (i.e., the allowable range of change set for each behavioral characteristic element) can be set based on the relative change (rate of change) compared to the user's reference point (reference behavioral characteristic information). That is, the device (10) can be configured (designed) to enable personalized judgment (search for changes in behavior) based on the change rate of the user's own reference.

[0231] In addition, the device (10) can be utilized for public and social purposes. For example, the device (10) can be utilized for public purposes by supporting installation during smartphone education or visits in conjunction with public health centers and local government elderly care programs, and through this process, users (e.g., the elderly) utilizing public health centers or care programs can be made able to use the service. The device (10) can be made available to users by installing the app at public institutions, and can provide installation guidance for the app to public health centers, welfare institutions, educational sites, etc.

[0232] Additionally, the device (10) may provide a life response-based self-search function (self-check function). In this regard, the device (10) may be equipped with, for example, a self-check module. In other words, the device (10) may further include a self-check module section (not shown) that provides a self-check function.

[0233] The self-check module above can provide a simple verification test function related to the user's usual reactions (usual behaviors) and digital behaviors (for example, providing questions for a verification test). That is, the self-check module can periodically (for example, at a period of 6 months, 1 year, etc.) provide pre-generated self-check test questions (test items) to the user terminal (20) for non-diagnostic purposes in order to perform a self-check test to identify information about the user's usual reactions and digital behaviors, and can enable the user to participate in the self-check test by inputting a response to the provided test questions.

[0234] In this case, the self-assessment module can provide rewards to users who have participated in the self-assessment test (i.e., entered responses to the test questions), thereby encouraging continuous participation in the self-assessment test. That is, the self-assessment module can provide a participatory reward function that provides rewards upon participation in the self-assessment test, which enables increased participation rates and long-term tracking, and induces voluntary and repetitive participation. In this case, local currency, points, etc., can be provided as rewards.

[0235] The self-check module can periodically provide a self-check test to the user terminal (20), thereby allowing the user to experience a self-check function based on their usual reaction trends at a 6-month or 1-year interval. This self-check test function (self-check function) does not have a medical diagnosis or discrimination function and can be configured (provided) as a non-medical device type behavioral response tool designed through expert consultation.

[0236] In other words, the self-assessment test items (test items) included in the self-assessment test can be generated (designed), for example, based on expert consultation, and the self-assessment test items may be referred to by other terms such as reactive test items. In addition, the self-assessment module may generate a lifestyle reference guide regarding changes in recurring behavioral tendencies, which is not for diagnostic purposes, based on medical institutions or expert consultation in the future, and provide it to the user. That is, the device (10) may generate and provide the lifestyle reference guide, which is an interpretation guideline for the flow of the user's behavioral changes, through consultation with medical staff in the future.

[0237] In summary, the self-monitoring module described above can periodically provide the user with a function to observe simple behavioral patterns based on the user's daily digital response tendencies (i.e., a self-monitoring function). If a pattern or change different from the usual pattern is repeated, the relevant information can be transmitted to a guardian or other relevant person with the user's consent. This self-monitoring function can be designed as a reference tool for recording and sharing the flow of changes in daily responses, rather than for medical diagnosis; it can be improved through expert consultation and designed to induce user participation through the linkage of rewards.

[0238] In addition, the self-inspection module can manage (control) that all data related to the self-inspection test is stored locally and is not shared externally without the user's explicit consent.

[0239] In addition, in describing the present invention, local (user terminal) execution was mentioned only in some components (e.g., a self-check module) within the device (10), but it is not limited thereto. The calculation configuration of the group average or the outlier processing procedure (cutoff, windsoring, etc.) described later may also be performed as much as possible on the local terminal, which is the user terminal (including the NPU). Furthermore, the data transmitted from the user terminal to the server, which is the device (10), may be, for example, refined statistical data rather than raw data. By transmitting such refined statistical data, the device (10) can ensure the protection of personal information and the minimization of delay.

[0240] In addition, regarding the medical device status and legal safety of the present invention, the following applies. The present invention (i.e., the device) is not intended for the diagnosis of disease or medical judgment, and is not classified as a diagnostic device or a therapeutic device under the Medical Devices Act. The device (10) identifies a relative change trend by comparing it with the user's usual digital behavior pattern, and provides the relevant information (i.e., information regarding the identified change trend) as a reference notification to a guardian, etc., which can serve as an opportunity for autonomous response to the user's life reactions.

[0241] The device (10) can be viewed as a non-medical, daily-based information delivery technology that deviates from the existing diagnosis-centered approach and, when a behavioral tendency different from usual that is placed in a blind spot appears based on the user's unconscious behavioral changes, conveys this to a guardian (family member), etc., thereby providing an opportunity for response at the level of guidance on cautious daily life reactions rather than diagnosis.

[0242] The device (10) can be described as having both social value and technical feasibility in the convergence area of ​​welfare / public service / AI healthcare. In addition, the device (10) has a structure that links with a guardian, and may provide a function to generate, for example, a report on a user's abnormal history and transmit it to the user or guardian.

[0243] In addition, in the above-described embodiment of the present invention, the device (10) generally provides examples only for cases where it searches for changes in a user's behavior based on a self-referential comparison structure that compares the user's reference behavior (reference behavior characteristic information) and current behavior (current behavior characteristic information), but this is merely an example to aid in understanding the present invention and is not limited thereto. As another example, the device (10) may search for changes in a user's behavior based on a group-based comparison structure that compares the user's current behavior (current behavior characteristic information) with the reference behavior of other users different from the user (in particular, group reference behavior characteristic information, which is the average normal pattern characteristic of other users). That is, the device (10) can explore (observe) changes in the user's behavior by using a self-referential comparison between the user's reference behavior and current behavior and / or a group-based comparison structure (i.e., a group-based comparison structure that performs a relative comparison with the average of a group of the same age group or similar characteristics).

[0244] At this time, the device (10) can perform an outlier processing procedure in the process of calculating the group average (i.e., the average of a group of similar characteristics). In relation to the outlier processing procedure, the device (10) can secure statistical stability by removing data corresponding to, for example, the upper / lower 10% range in the process of calculating group standard behavioral characteristic information using a cut-off method. At this time, the percentage range may vary depending on the implementation environment and purpose of the present invention, and may be adjusted, for example, to a range of about 5% to 15%. Meanwhile, separately from the cut-off processing, a portion of the total data (for example, about 1% to 5%) is transmitted to the server, which is the device (10), and can be used for readjusting or correcting Winsorizing values. Through this, the device (10) can secure immediate stability locally while enabling long-term / dynamic correction on the server.

[0245] In other words, the device (10) can explore (observe) whether a change in a user's behavior deviates from a statistically significant level through a group-based comparison, which performs not only a self-referential comparison but also a comparison with the average pattern (group reference behavioral characteristic information) of multiple users (other users) who are of the same age group, gender, or have similar usage patterns as the user. That is, the device (10) can explore a change in the user's behavior (behavioral rhythm) by observing whether the user's current behavior deviates from the group reference behavioral characteristic information statistically through a group-based comparison. In other words, when performing a group-based comparison, the device (10) can perform a comparison of the group reference behavioral characteristic information based on age, gender, or similar characteristics. In this case, the term 'group' considered in the present invention is not limited to static criteria (age, gender, etc.) but can be defined as a group of users that is dynamically reconstructed to show similar patterns (e.g., typing speed reduction rate, increase in error rate, etc.) over time. That is, the groups considered in the device (10) may include not only statically configured groups but also dynamically configured groups.

[0246] In addition, the device (10) may also explore changes in user behavior by utilizing, for example, user clustering based on artificial intelligence (AI) models, anomaly detection algorithm based on group statistics, etc.

[0247] In addition, based on the technology of the device (10), the present invention enables the following brand expansion (expansion into derivative services). Future service lineups and expansion models (expansion examples) based on the technology of the present invention may include, for example, i) 'Mom, Why Are You Like That', a service centered on voice-based behavior change detection functions (i.e., a service that detects behavior based on voice), ii) 'Mom's Word', a tool for detecting user emotional changes and family communication, iii) 'Onki for Caregiver', a B2B2C monitoring solution for caregivers, and iv) 'Mom's Homemade Taste', a recipe content platform that allows recording and sharing parents' cooking in voice and text. Here, 'Mom's Word' can be described as an emotional legacy tool that archives a mother's tone of voice or expressions. At this time, the brand name (service name) can be set and changed in various ways by the administrator as needed, and can be further expanded in addition to the aforementioned services. Furthermore, the aforementioned expansion brands can be described as application services derived based on the technology of the device (10).

[0248] Users of the device (10) may include all smartphone users, not limited to a specific age group. The device (10) may be a non-medical emotional notification device that is not intended for the diagnosis or classification of diseases, but rather delivers a notification for reference based on changes in daily behavior to a guardian or registered account.

[0249] As described above, the acquisition unit (12) can acquire (collect) digital behavior data of a user (i.e., typing data, voice data, activity data, etc.) from a user terminal (20), and the digital behavior data acquired by the acquisition unit (12) may be referred to as passive data.

[0250] The above passive data refers to data automatically collected by the system without the user separately recognizing or intending it. For example, it may include data regarding typing input speed (FT / HT), error rate, frequency of backspace usage, and frequency of entry into clock or calendar apps, measured during the use of a smartphone keyboard (i.e., a third-party keyboard installed on the user's terminal). Such passive data can be automatically collected (i.e., involuntary automatic collection during normal use) simply by the user using the smartphone (user terminal) as usual. In this case, regarding typing input speed, FT (Flight Time) refers to the interval between keys, and HT (Hold Time) refers to the duration of the key press.

[0251] The acquisition unit (12) can acquire (collect) active data from the user terminal (20) in addition to the digital behavior data, which is passive data described above. At this time, to acquire the active data, the control unit (17) can provide (present) today's sentence to the user terminal. By providing today's sentence, the control unit (17) can induce the user to input (type and input) today's sentence. When the user inputs today's sentence provided (presented), the acquisition unit (12) can acquire sentence input data generated by inputting today's sentence (i.e., sentence input data, which is the user's input data corresponding to today's sentence) as the active data.

[0252] The above active data may refer to data collected only when the system induces a specific action (i.e., an action of having the user input / type the presented sentence of the day) and the user performs it voluntarily. It may include data regarding typing speed (FT / HT) or error rate collected through the user directly typing or speaking the presented sentence of the day. Such active data requires user participation and may be collected only when the user consciously inputs it; this has the characteristic of being advantageous for quantitative analysis as the system can repeatedly collect data in a certain format.

[0253] The device (10) can detect changes in behavior that differ from the user's usual behavior by considering passive data and active data obtained (collected) from the user terminal (20).

[0254] In the example of the present invention described above, the focus was on explaining the technology in which the device (10) passively collects and analyzes the user's digital behavior data to explore behavioral changes different from the user's usual behavior. In another embodiment of the present invention below, the technology in which the device (10) actively collects not only digital behavior data but also active data (i.e., sentence input data) from the user through active input will be explained in detail. That is, in addition to the aforementioned passive collection and analysis technology of digital behavior, the device (10) aims to provide a more expanded technology for user intervention and inducing emotional connection by additionally implementing the structure of active input (e.g., input for 'sentence of the day'), methods of utilizing it, and reward-based participation induction methods.

[0255] The device (10) can passively collect (acquire) digital behavior data, which is the user's smartphone usage data, and additionally collect standardized input data such as 'sentence of the day' that the user actively inputs (i.e., sentence input data generated by the user typing the given sentence of the day) as active data, and measure the user's daily change state through the analysis of the active data, and as a result, based on the analysis of the passive data and the active data, it can automatically detect abnormal signs or emotional change states of the user (i.e., detect changes in behavior different from the user's usual behavior).

[0256] Specifically, the acquisition unit (12) can acquire digital behavior data of the user from the user terminal (20) for acquiring passive data.

[0257] Additionally, for the acquisition of active data, the control unit (17) may provide today's sentence to the user terminal (20). Here, today's sentence may refer to a sentence (today's sentence) that is predefined (set) by an administrator and provided to the user every day. The control unit (17) may provide today's sentence, which is predefined by the administrator, to the user terminal (20) at a predefined time (e.g., 10:00 AM) every day. The today's sentence may be provided as a different sentence every day. The control unit (17) may provide today's sentence to the user terminal (20) to induce the user to input (type input) today's sentence provided (presented).

[0258] The input structure of today's sentence for collecting active data may be as follows. For example, when the control unit (17) provides today's sentence to a user terminal (20), it may provide an input screen for inputting today's sentence in the form of a gamified UI, and may provide visual feedback in the form of, for example, highlighting the word corresponding to the typo in red in real time when a typo occurs during the input of today's sentence.

[0259] The change detection unit (15) can detect the change (i.e., a change in behavior different from the user's usual behavior) by further considering the analysis information of the user's sentence input data when the user inputs the sentence of the day, which is generated by the user inputting the sentence of the day, is collected from the user terminal (20). Accordingly, the acquisition unit (12) can further collect the user's sentence input data from the user terminal (20).

[0260] The change detection unit (15) can automatically assign (tag) status tags to data of preset indicators included in the acquired passive data and active data after passive data (digital behavior data) and active data (sentence input data) are acquired from the user terminal (20). Here, the preset indicators are indicators for detecting (detecting) changes in the user's behavior (changes in behavior different from usual behavior), and may be referred to by other terms such as behavior change detection indicators.

[0261] The above-mentioned preset indicators may include, for example, a first indicator corresponding to typing speed (typing input speed) derivable through the analysis of passive data, a second indicator corresponding to the error rate derivable through the analysis of passive data, a third indicator corresponding to the entry frequency (access frequency) for a specific app (e.g., clock app, calendar app) derivable through the analysis of activity data within passive data, a fourth indicator corresponding to typing speed derivable through the analysis of active data, and a fifth indicator corresponding to the error rate derivable through the analysis of active data.

[0262] Here, the fourth and fifth indicators may refer to the typing speed and error rate when the user inputs the sentence of the day provided (presented) to the user (i.e., the typing speed and error rate when inputting the sentence of the day). The typing speed considered in the fourth indicator may be referred to by other terms such as the speed of inputting a standard sentence, and the error rate considered in the fifth indicator may be referred to by other terms such as the error rate when inputting a standard sentence. In addition, the typing speed considered in the first and fourth indicators may be data measured based on the interval between keys, FT (Flight Time), and the duration of key press, HT (Hold Time).

[0263] When the change detection unit (15) automatically assigns a status tag to the data (measurement value) of the preset indicators, the status tag can be automatically assigned based on the relative distribution within a homogeneous group (e.g., a group of the same age group or a group of similar behavior). When assigning the status tag, the change detection unit (15) can assign one English letter sequentially in alphabetical order to each of the data (measurement value) of the preset indicators.

[0264] At this time, the status tag may be configured (prepared) based on the level (percentage range) within a homogeneous group corresponding to the degree of change of each measurement value of the preset indicators (e.g., the first to fifth indicators) relative to the user's personal baseline. For example, the status tag may be displayed in uppercase English letters when it corresponds to a preset top percentage range within a homogeneous group (e.g., a range of less than the top 10%), and in lowercase English letters when it corresponds to a preset top percentage range within a homogeneous group (e.g., a range of more than 10% and less than 30%).

[0265] That is, if an uppercase English letter is assigned as a status tag to the measurement values ​​of the pre-set indicators, it may mean that the user falls within the top 10% range within the homogeneous group (i.e., a user who has shown a sudden change in behavior). On the other hand, if a lowercase English letter is assigned as a status tag to the measurement values ​​of the pre-set indicators, it may mean that the user falls within the top 10% or more but less than 30% range within the homogeneous group (i.e., a user who has shown a less severe change in behavior compared to the aforementioned user).

[0266] In addition, in the present invention, when assigning status tags based on percentage intervals within a group (e.g., top 5%, 10%, etc.), the assigned status tags are not assigned based simply on absolute value comparisons, but may be assigned based on data evaluating 'how relatively rapid an individual's change is compared to the group average change' (relative change detection concept).

[0267] For example, the change detection unit (15) may assign a status tag A to the first indicator when the level of decrease in the user's first indicator data (typing speed) falls within the top 5% of the homogeneous group, ii) assign a status tag b to the second indicator when the level of increase in the user's second indicator data (typing error rate) falls within the top 20% of the homogeneous group, iii) assign a status tag c to the third indicator when the level of increase in the user's third indicator data (frequency of entry into a specific app) falls within the top 15% of the homogeneous group, iv) assign a status tag D to the fourth indicator when the level of decrease in the user's fourth indicator data (typing speed when inputting today's sentence) falls within the top 8% of the homogeneous group, and v) assign a status tag E to the fifth indicator when the level of increase in the user's fifth indicator data (typing error rate when inputting today's sentence) falls within the top 9%. Consequently, the change detection unit (15) can assign the status tag, for example, 'AbcDE', to the data of the user's preset indicators.

[0268] According to this, the state tag assigned by the change detection unit (15) may be composed of a combination of uppercase and lowercase letters of the alphabet, such as 'AbcDE', 'ABCdE', etc., and may be designed to enable complex state interpretation. That is, the change detection unit (15) can detect (detect) a change in the user's behavior (a level of changed behavior different from the user's usual behavior, etc.) based on the state tag assigned (tagged) to the user's pre-set indicator data.

[0269] At this time, the names (e.g., uppercase English letters such as A, B, C, etc. or lowercase English letters such as a, b, c, etc.) and numbers (quantities) of the status tags assigned to each indicator (e.g., the first to fifth indicators, etc.) and the names and numbers (quantities) of each indicator are merely examples to aid in understanding the present invention and are not limited thereto; they can be set and changed in various ways by the administrator, and can be expanded to E, F, G, etc., if various digital behavior indicators are considered in the future. Furthermore, in describing the present invention, the configuration method (types or number) of each indicator, reference values ​​(e.g., the top percentage range, which is less than the top 10%, and the top percentage range, which is between the top 10% and less than 30%), and the classification range of age groups for distinguishing homogeneous affiliated groups are not limited only to the examples described above, but can be set and changed in various ways by the administrator and can be changed in various ways during the verification or advancement process. In addition, while the example described above illustrates homogeneous groups classified based on age, it is not limited to this, and homogeneous groups can be flexibly formed based on criteria such as gender and region as well as age.

[0270] In the above description, the data (measurement value) regarding the fourth indicator and the fifth indicator may be data that can be collected (obtained) by providing the 'sentence of the day' to the user terminal (20) every day and then inducing the user to input the provided 'sentence of the day' every day. In particular, the data (measurement value) regarding the fourth indicator and the fifth indicator may be data that can be obtained through the analysis of the user's sentence input data corresponding to the sentence of the day, and this may be referred to by other terms such as daily typing pattern data for the sentence of the day.

[0271] Additionally, the change detection unit (15) can calculate the relative position (percentile, rank, etc.) of the user within the distribution by comparing the rate of change of the user’s behavior indicator, for example, with the distribution of the rate of change of a group classified by the same age and gender, and can provide the user with a risk alert, a prevention alert, or positive feedback based on the calculated result. In this regard, for example, if the user’s input speed decrease rate over the past 4 weeks is measured as -8%, the change detection unit (15) can confirm that it corresponds to the 65th percentile by comparing it with the distribution of the rate of change of the same age / gender group, and accordingly, can display (provide) a message on the screen stating "a decrease corresponding to the top 35% of peers." According to this, the present invention (the device (10)) can provide not only a simple comparison with the group average but also percentile and rank calculation results based on the distribution within the group to the user and guardians.

[0272] In the present invention, the acquisition unit (12) can continuously perform collection (acquisition) of passive data in parallel when collecting active data. That is, the acquisition unit (12) can continuously perform collection of active data and collection of passive data in parallel.

[0273] The control unit (17) can provide a reward to the user terminal when the user inputs the sentence input data for today's sentence. The control unit (17) can provide a reward to the user when the user inputs (types) the sentence of the day provided (presented) to the user.

[0274] The control unit (17) can provide a reward to the user terminal (20) when the user inputs a designated (presented) sentence (today's sentence), thereby inducing behavior-based repetition. Here, the reward can be provided not only as a basic input reward but also additionally according to cumulative conditions, and can be designed (set) according to various policies, such as an additional reward for inputting for 10 consecutive days or an additional reward for perfect attendance for a month. In addition, the reward can be provided as points, local currency, or other cash-like values.

[0275] That is, for example, the control unit (17) can provide a reward worth 50 won for every input made once a day. In addition, the control unit (17) can provide additional rewards for 10 cumulative days, 20 cumulative days, and monthly perfect attendance, and in the case of additional rewards, a total of up to 3,900 won worth of rewards can be provided per month. At this time, the rewards are, in principle, provided by local government regional gift certificates or external entities, but can be internally provided as substitutes such as mobile gift certificates or points. The amount and structure (provision form) of the rewards provided by the control unit (17) can be designed to be flexible and can be set and changed in various ways by the manager.

[0276] In addition, the ‘sentence of the day’ provided by the control unit (17) can be configured (prepared) in the form of, for example, a policy message (e.g., a heat wave warning), an emotional message (e.g., you are not alone), public information (e.g., vaccination notice). The user’s input to the presented sentence of the day (i.e., input of sentence input data) itself can be used as a standardized criterion for analyzing the user’s state.

[0277] Additionally, the control unit (17) can enable the user to input the sentence of the day presented (provided) by voice input as well as by typing input. That is, the sentence input data may be text data entered by typing input or voice data entered by voice input. The control unit (17) may support a voice input function for the sentence of the day in this way. The control unit (17) can replace input by speaking a predetermined sentence (the sentence of the day presented) in voice when typing is difficult for the user. At this time, when voice data regarding the sentence of the day is input by the user, the voice data may be expanded in conjunction with an AI (artificial intelligence) model capable of future sentiment analysis and analysis of voice speed / tone changes. That is, for example, when the voice data is input, the device (10) can perform sentiment analysis and analysis of voice speed / tone changes, etc., using an AI model on the input voice data, and can explore changes in the user's behavior by further considering (utilizing) the analyzed information.

[0278] According to this, the control unit (17) can operate a reward system (i.e., a reward-based participation inducement system) that provides a sentence of the day to the user every day and provides a reward whenever the user inputs the provided sentence of the day. The control unit (17) includes a reward structure that can induce repeated use of the device (10) by the user through the provision of the reward system.

[0279] Additionally, the control unit (17) may provide an emotional notification function based on QR generation, for example (i.e., a QR-based emotional notification function). In this regard, the control unit (17) may control the generation of a QR code and display it on the screen of the user terminal (20) when a predetermined standard is satisfied (for example, when the assignment (tagging) of the aforementioned status tag is completed, or when the detection time of the aforementioned non-response persists for longer than a preset time). At this time, the QR code displayed on the user terminal (20) may be referred to as a QR-based emotional notification (emotional notification information).

[0280] At this time, the control unit (17) can enable the user to present the user terminal (20) with the QR code displayed to a relevant institution such as a public health center, hospital, or welfare center. Subsequently, the control unit (17) can enable a person in charge of the relevant institution to scan the QR code displayed on the user terminal (20) to decode and verify information such as the reason why the QR-based emotional notification was provided to the user terminal (20) through the scan, the user's status code (i.e., information on the status tag assigned to the user), the time of the QR-based emotional notification, and whether the user has entered today's sentence. According to this, the control unit (17) can induce a response from the guardian or relevant institution regarding the user's emotional state or abnormal signs without the guardian installing a separate app (this app), and the act of scanning the QR code itself can be recorded as an expression and practice of interest in the user. This device (10) can provide a structure that allows various relevant parties (those related to the user), such as guardians, public health centers, and welfare institutions, to access user-related information and induce emotional intervention.

[0281] Additionally, the control unit (17) can provide a dashboard for administrators and an institutional interface. At this time, the screen of the administrator dashboard can be designed in a structure that allows not only the technical operator of the system but also personnel in charge of related institutions, such as public health centers, welfare centers, and local government practitioners, to use it. The control unit (17) can enable personnel in charge of related institutions to check information such as the decoding result of the QR code, the user's status, the input history of today's sentence by the user, and the trend of recent changes in the user's behavior through the screen of the administrator dashboard, and can utilize this information for administrative intervention and service linkage judgments. Additionally, the control unit (17) can record and manage logs such as the verification history and intervention history of personnel in related institutions in the database unit (13) within the device (10).

[0282] Additionally, the device (10) may provide a guardian interface, and in this regard, the notification management unit (16) may transmit notification information, which is an emotional notification, to the guardian terminal as previously mentioned. At this time, the notification management unit (16) may, for example, allow the guardian to access the app (dedicated app) or other notification channels through the guardian terminal to receive and verify information regarding the status of the user being protected as a notification (i.e., notification information). Here, the notification information (notification message) provided to the guardian terminal may be configured to include, for example, indirect information regarding the user's recent reaction patterns or whether structured input is present (i.e., whether input regarding today's sentence is present), rather than including specific actions or location information of the user. According to this, the device (10) may enable the guardian to check information such as whether the user has entered today's sentence and the time of receiving the notification in a summarized form through the app or the notification information, thereby inducing an emotional response and promoting repetitive interest behavior.

[0283] Additionally, the control unit (17) may provide emotional feedback to, for example, a user terminal (20) and / or a guardian terminal. In this regard, when the control unit (17) detects that an interest event regarding the user has occurred, it may provide emotional feedback containing non-medical emotional phrases to the user terminal (20) and / or the guardian terminal. Here, the interest event may refer to an event corresponding to an act in which someone takes an interest in the user, and may refer to, for example, an act of a relevant agency scanning a QR code displayed on the user terminal (20), or an act of a guardian accessing the app through the guardian terminal out of curiosity about the user's condition. The non-medical emotional phrases included in the emotional feedback may, for example, be "Thank you for turning on the ONKI app," "Your family is thinking of you," etc. Here, ONKI may, for example, refer to the service name of the app.

[0284] The control unit (17) can control the tone or content of the message (non-medical emotional phrase) to change according to the user's state when providing the emotional feedback, and can provide it in a customized form according to the user's response history or the guardian's participation history. The device (10) may be configured (designed) to have a structure that basically transmits a notification (notification information) to the guardian, but is not limited thereto, and may be configured (designed) to provide a notification directly to the user or provide a notification to the guardian or the user in parallel when necessary.

[0285] In addition, the device (10) can be advanced not only to the functions described above, but also to an AI-based prediction model, behavior clustering, and real-time anomaly detection algorithm, and can be developed into a customized notification message recommendation system (e.g., message optimization based on guardian behavior history).

[0286] As described above, the device (10) can induce a user to input today's sentence, provide a reward for inputting today's sentence, and provide a QR-based emotional notification. In particular, the device (10) can not only acquire existing passive typing-based quantitative indicators (e.g., input speed, typing error rate, backspace rate, etc.) but also induce a user to actively input a predetermined sentence (i.e., today's sentence) to acquire sentence input data for that sentence as active data. The device (10) can increase the accuracy of repeated measurements by providing a reward based on the input result of today's sentence, and can send a notification to a guardian, etc., if the user's input does not continue for a certain period (i.e., if no input for today's sentence is made for a certain period).

[0287] The present device (10) may be referred to by other terms such as a digital behavior-based behavior change detection device, as well as a digital behavior-based user state detection device for inducing emotional responses and arousing interest. The present device (10) detects emotional changes in the user's state based on digital behavior data, such as the daily input and touch patterns of a smartphone user, and accordingly induces intervention of interest from guardians and public institutions.

[0288] In modern society, digitally vulnerable individuals, such as the elderly and those living alone, are prone to experiencing social isolation or emotional disconnection as they spend more time alone. While some care systems exist for them, they require active participation or focus on medical intervention, which limits their ability to induce emotional connection or interest in daily life. To address this problem, the present device (10) provides a structure that detects signs of behavioral change based on the user's smartphone usage patterns and induces non-medical interest intervention accordingly.

[0289] The device (10) can collect (acquire) the user's passive and active data and automatically assign status tags based on relative distribution to the collected data. Subsequently, the device (10) can generate a QR code and display it on the user terminal (20) when the above-described certain criteria (specified conditions) are satisfied, and can control the generated QR code to include user-related information (including user status information). Subsequently, when, for example, an administrator or a person in charge of a relevant agency scans the QR code displayed on the user terminal (20), the device (10) can decrypt the scanned QR code to enable inquiry (confirmation) of the user's status and whether structured input has been entered (i.e., whether today's sentence has been entered), and based on this, induce intervention regarding the user's status.

[0290] In addition, the device (10) can provide rewards for the user's repeated participation in the problem of the day provided to the user every day, and the reward conditions and means of compensation may be flexibly designed and changed. The device (10) can evaluate the user's state or provide rewards by utilizing the user's structured sentence input (sentence of the day). At this time, the structured input (i.e., input for the sentence of the day) may be replaced not only by a typing input method but also by a voice input method, and the voice input may be used for subsequent analysis or user state estimation. Furthermore, the device (10) and the system may be configured to have a structure that can be expanded in the future to include state prediction, notification optimization, and user clustering functions through an AI-based model.

[0291] In other words, the device (10) collects digital behavior data of the user, analyzes passive and active data including input speed, typing error rate, frequency of entry to the clock and calendar, and whether the 'sentence of the day' is performed, evaluates the user's status according to a predefined ABCD classification system (i.e., criteria for assigning predefined status tags), and can provide a notification (emotional notification, status notification) to a guardian, welfare agency, or the user themselves when a certain standard is exceeded (satisfied), thereby detecting changes in the user's digital behavior.

[0292] At this time, the device (10) can provide a reward to the user based on whether the user inputs today's sentence and accuracy, etc., and can provide additional rewards when the cumulative number of today's sentence inputs or a predetermined condition is met, thereby inducing reward-based behavior.

[0293] Additionally, the device (10) may be configured to include a self-notification module that induces a change in behavior by displaying a self-warning message in the form of a pop-up, vibration, voice guidance, etc., to the user terminal (20) when it is detected that the user's state deviates from a predetermined standard state. Such a self-notification module may refer to a configuration included in the notification management unit (16) described above.

[0294] Additionally, the device (10) can generate a QR code and display it on the screen of a user terminal (20) when a predetermined standard is satisfied. The QR code displayed on the user terminal (20) may include information such as the user's status code, input history for today's sentence, and reward status. The device (10) can enable a person in charge of a relevant institution, such as a welfare institution or a public health center, to check the user's status by scanning the QR code with their terminal. That is, the device (10) may be configured to include a QR-based status linkage module that allows the user's status to be checked, for example, when the QR code is scanned from the person in charge's terminal.

[0295] In addition, the device (10) can enable a guardian to check the user's status and receive notifications via text message, email, browser link, etc., without installing a separate app, and the device (10) can provide a non-identifiable guardian notification function.

[0296] In addition, the device (10) can enable the formation of an emotional connection by allowing expressions of interest or feedback messages entered by a guardian or relevant agency to be displayed as emotional messages on a user terminal (20) (i.e., by providing emotional feedback including non-medical emotional phrases). The device (10) may be configured to include, for example, a feedback interaction module capable of providing emotional messages, and the feedback interaction module may be a component included in, for example, a notification management unit (16).

[0297] In addition, the device (10) can learn, for example, collected passive data and active data using an artificial intelligence model (machine learning prediction model), and can predict the possibility of deterioration of the user's condition (behavioral condition) using such learned artificial intelligence model and provide a pre-warning alarm to a guardian or relevant agency (i.e., provide an abnormality detection alarm function).

[0298] In addition, the aforementioned predefined ABCD classification system described above may refer to the criteria for assigning the previously described predefined status tags. The predefined ABCD classification system may be composed, for example, of [A: rate of decrease in input speed (group standard), B: change in typo rate, C: increase in frequency of clock / calendar entry, D: decrease in today's sentence input speed (absolute or relative standard), etc.], and the device (10) can determine the user's status according to each indicator using this predefined ABCD classification system.

[0299] In addition, the device (10) can enable the user to input today's sentence using either a voice input method or a text input method (i.e., a text input method by typing) when inputting today's sentence presented by the user, and can analyze the input speed and accuracy of today's sentence according to the input method.

[0300] In addition, the device (10) can quantify the frequency of feedback and the number of times interest is expressed by the guardian and use this as an emotional intervention index, and can control (set) additional intervention or system intervention conditions for the user according to the index (emotional intervention index).

[0301] In addition, the device (10) can be configured to transmit the user's status to a guardian or relevant agency using various communication means such as NFC, BLE, and CAPTCHA, in addition to the aforementioned QR code, and thus can have a complex linkage structure. Furthermore, the device (10) can provide a dashboard screen that allows relevant agencies, such as welfare institutions or public health centers, to check the status, change history, feedback details, and reward status of each user, and through this dashboard screen, the guardian or relevant agency can comprehensively manage and check the user's status.

[0302] In addition, the artificial intelligence model (machine learning prediction model, machine learning model) considered in the device (10) can be configured to form clusters within a user group (in particular, forming multiple homogeneous membership groups based on criteria such as age, gender, region, etc.) and detect abnormal patterns based on the difference from the mean within the clusters, thereby enabling early identification of the user's risk status. The device (100) can detect abnormalities in users based on clusters using the artificial intelligence model.

[0303] This device (10) can provide a state detection and response induction structure designed around an 'emotional connection' between the user, guardian, and relevant agency, distinct from a medical alert system. The device (10) can be configured to have a structure that can induce the attention of guardians and relevant agencies even from non-responsive users without separate operation. Additionally, the device (10) can facilitate linkage with public welfare / digital care systems through the formation of an emotional feedback loop between relevant agencies and the user.

[0304] In addition, the device (10) can improve the normality and analysis accuracy of the data through active input actions such as 'sentence of the day', and can also provide opportunities for emotional communication. Furthermore, the device (10) can be expanded into an AI-based precision classification and prediction system through future advancements.

[0305] In addition, the above description describes the process of detecting behavioral changes primarily based on typing data when describing the device (10). This is merely an example to aid in understanding the invention and is not limited thereto. The technology proposed by the device (10) can apply the same principle not only to the aforementioned digital behavioral data (user typing data, voice data, activity data) but also to continuous behavioral data such as walking data (e.g., walking interval, steps per minute, etc.). That is, the device (10) can be extended and applied to various behavioral indicators such as walking data as well as the aforementioned digital behavioral data.

[0306] In addition, in describing the present invention (i.e., the system and the device (10)), all source data considered in the present invention (e.g., typing data, voice data, activity data, walking data, etc.) are primarily analyzed in a local environment within the user terminal (20), and after the analysis is completed, the source data may be immediately destroyed from volatile memory. That is, raw data such as typing input streams, original voice waveform data, location logs, or app usage logs are not transmitted outside the terminal (user terminal (20)), and may be configured to be automatically deleted after feature extraction is performed in a temporary memory area within the terminal.

[0307] Information transmitted from the user terminal (20) to the server, the device (10), may be limited to non-identifiable statistical feature values, and the statistical feature values ​​may be numerical characteristic information that cannot directly identify an individual, such as average value, standard deviation, coefficient of variation, time series slope, and rate of change indicator.

[0308] According to this, the present invention can technically implement the 'Privacy by Design' concept and comply with the principle of minimal collection of personal information by adopting a structure that, in principle, excludes the external transmission of source data and transmits only anonymized feature information to the server. Furthermore, the structure can be designed so that the server cannot reconstruct an individual's specific input content, voice content, or location path, thereby providing the effect of structurally guaranteeing the protection of user privacy.

[0309] Below, we will briefly examine the operation flow of the present invention based on the details described above.

[0310] FIG. 3 is an operation flowchart of a digital behavior-based behavior change detection method according to an embodiment of the present invention.

[0311] The digital behavior-based behavior change detection method illustrated in FIG. 3 can be performed by the device (10) described above. Therefore, even if the content described below is omitted, the description of the device (10) can be equally applied to the description of the digital behavior-based behavior change detection method.

[0312] Referring to FIG. 3, in step S11, the acquisition unit can acquire digital behavior data of the user from the user terminal.

[0313] Next, in step S12, the standard setting unit can set the user's standard behavior characteristic information through the analysis of the digital behavior data.

[0314] Next, in step S13, the change detection unit can detect changes in behavior that differ from the user's usual behavior within the additional digital behavior data by analyzing the additional digital behavior data additionally acquired after the setting of the reference behavior characteristic information.

[0315] In the description above, steps S11 to S13 may be further divided into additional steps or combined into fewer steps, depending on an embodiment of the present invention. Additionally, some steps may be omitted as necessary, and the order between steps may be changed.

[0316] A digital behavior-based behavior change detection method according to one embodiment of the present invention may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either individually or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the present invention, or they may be those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. The hardware device may be configured to operate as one or more software modules to perform the operation of the present invention, and vice versa.

[0317] In addition, the aforementioned digital behavior-based method for detecting behavioral changes can also be implemented in the form of a computer program or application executed by a computer and stored on a recording medium.

[0318] The foregoing description of the present invention is for illustrative purposes only, and those skilled in the art will understand that other specific forms can be easily modified without altering the technical spirit or essential features of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single unit may be implemented in a distributed manner, and components described as distributed may likewise be implemented in a combined form.

[0319] The scope of the present invention is defined by the claims set forth below rather than by the detailed description above, and all modifications or variations derived from the meaning and scope of the claims and equivalent concepts thereof should be interpreted as being included within the scope of the present invention.

Claims

1. An acquisition unit that acquires digital behavior data of a user from a basic input interface of an application or operating system installed on a user terminal; A standard setting unit for setting user standard behavioral characteristic information through the analysis of the above digital behavioral data; and A change detection unit that searches for changes in behavior different from the user's usual behavior within the additional digital behavior data through the analysis of additional digital behavior data additionally acquired after the setting of the above standard behavior characteristic information, A digital behavior-based behavior change detection device including 2. In Paragraph 1, The above change detection unit is, Explore changes in the behavior by comparing the current behavioral characteristic information derived from the additional digital behavioral data with the reference behavioral characteristic information, A digital behavior-based behavior change detection device that explores behavioral changes by analyzing the time-series trend and speed of change of user-specific data.

3. In Paragraph 1, The above digital behavior data includes user typing data, and A digital behavior-based behavior change detection device, wherein the standard setting unit derives a typing pattern regarding at least one of typing speed, rhythm, typing position, number of typos, and number of deletion inputs through the analysis of the typing data, and sets it as typing-related standard behavioral characteristic information.

4. In Paragraph 1, The above digital behavior data includes the user's voice data, and A digital behavior-based behavior change detection device, wherein the above-mentioned standard setting unit derives a voice pattern regarding at least one of the user's speech rhythm, prosodic features, and disfluency features through the analysis of the above-mentioned voice data and sets it as voice-related standard behavior characteristic information.

5. In Paragraph 1, The above digital behavior data includes user activity data, and A digital behavior-based behavior change detection device, wherein the above-mentioned standard setting unit derives at least one of an app usage pattern, an inactive state pattern, and a location pattern through the analysis of the above-mentioned activity data and sets it as activity-related standard behavior characteristic information.

6. In Paragraph 1, A notification management unit that transmits notification information indicating that there is a change in the user's behavior when the above-mentioned change in behavior is detected, A digital behavior-based behavior change detection device further comprising 7. A digital behavior-based behavior change detection system comprising the digital behavior-based behavior change detection device of claim 1, A user terminal that generates digital behavior data; and A digital behavior-based behavior change detection device that detects behavior changes by receiving digital behavior data or feature information generated based on digital behavior data from the above user terminal, A digital behavior-based behavior change detection system including 8. A method for detecting a change in behavior based on digital behavior using a digital behavior-based behavior change detection device of claim 1, In the acquisition unit, a step of acquiring digital behavior data of a user from a user terminal; In the standard setting unit, a step of setting user standard behavior characteristic information through analysis of the digital behavior data; and In the change detection unit, a step of detecting changes in behavior different from the user's usual behavior within the additional digital behavior data by analyzing additional digital behavior data additionally acquired after the setting of the reference behavior characteristic information, A digital behavior-based method for detecting behavioral changes including 9. A computer-readable recording medium storing a program for executing the method of paragraph 8 on a computer.

Images