Method and apparatus for real-time frailty index prediction based on biometric and lifelog data

Abstract

Provided are a method and apparatus for real-time frailty index prediction based on biometric and lifelog data. The method for real-time frailty index prediction, according to an embodiment of the present invention, may comprise the steps of: collecting and processing biometric and lifelog data; generating input data by extracting characteristics from the processed biometric and lifelog data; and inputting the generated input data into a frailty prediction artificial intelligence model and receiving an output of a frailty index prediction value and a key factor for improving a frailty index.

Description

Method and device for predicting real-time frailty indicators based on biometric information and lifelog data

[0001] The present invention relates to a method and apparatus for predicting real-time frailty indicators based on biometric information and lifelog data, and more specifically, to a method and apparatus for predicting real-time frailty indicators based on biometric information and lifelog data that processes biometric information and lifelog data measured in real time through a wearable device, predicts frailty indicators in real time through an artificial intelligence model, and suggests personalized optimal activities based on the prediction results.

[0002] With the advancement of wearable device technology, there are increasing attempts to utilize real-time biometric data in the field of digital healthcare. Biometric data measured by sensors embedded in wearable devices is primarily used to monitor heart rate to assess physical condition or to determine the quality of sleep. Despite these advancements, there is a problem in that there are limitations to evaluating aging and frailty, such as indicators of frailty, solely based on biometric data.

[0003] The method for calculating frailty indicators is still based on offline questionnaires administered by medical professionals. This existing approach has a problem in that it cannot accurately calculate frailty indicators due to subjective responses, and consequently, it fails to accurately suggest optimal, personalized activities based on objective criteria.

[0004] The present invention aims to solve the aforementioned problems by providing a method and apparatus for predicting real-time frailty indicators based on biometric information and lifelog data, which can predict frailty indicators in real time through an artificial intelligence model based on objective biometric information and lifelog data.

[0005] A method for predicting real-time frailty indicators based on bio-information and life-log data according to one embodiment of the present invention for achieving the above objective may include the steps of: collecting and processing bio-information and life-log data; extracting characteristics from the processed bio-information and life-log data to generate input data; and inputting the generated input data into an artificial intelligence model for frailty prediction to receive outputs of frailty indicator prediction values ​​and key factors for improving frailty indicators.

[0006] And the above-mentioned step of collecting and processing can perform processing to add missing values ​​to the collected biometric information and lifelog data using the mean and standard deviation extracted from the previously stored biometric information and lifelog data.

[0007] In addition, the step of collecting and processing the above may set health checkup data received from an external server as reference data and perform processing to correct the collected biometric information using the set reference data.

[0008] And the step of generating the input data may involve extracting time-series characteristics from the processed biometric information, extracting characteristics regarding whether there is a change, the amount of change, and the direction of change from the processed lifelog data, and integrating the extracted characteristics to generate the input data.

[0009] In addition, the output receiving step may input the input data into an artificial intelligence model trained using an LSTM algorithm to output a predicted value of the frailty index, and determine which of the input data needs to be improved to lower the frailty index value, thereby determining a key factor for improving the frailty index.

[0010] Meanwhile, a real-time frailty indicator prediction device based on biometric information and lifelog data according to one embodiment of the present invention for achieving the above objective includes a communication unit that communicates with an external device, a database that stores data, and a processor that controls the frailty indicator prediction device. The processor controls the communication unit to receive biometric information and lifelog data collected from a wearable device, processes the received biometric information and lifelog data, extracts characteristics from the processed biometric information and lifelog data to generate input data, and inputs the generated input data into a frailty prediction artificial intelligence model to output a frailty indicator prediction value and a key factor for improving the frailty indicator, and stores them in the database.

[0011] And the processor can perform processing to add missing values ​​to the collected biometric information and lifelog data using the mean and standard deviation extracted from the previously stored biometric information and lifelog data.

[0012] In addition, the processor can control the communication unit to receive health checkup data from an external server, set the received health checkup data as reference data, and perform processing to correct the collected biometric information using the set reference data.

[0013] And the processor can extract time-series characteristics from the processed biometric information, extract characteristics regarding whether there is a change, the amount of change, and the direction of change from the processed lifelog data, and integrate the extracted characteristics to generate the input data.

[0014] In addition, the processor can input the input data into an artificial intelligence model trained using an LSTM algorithm to output a predicted value of the frailty index, and determine which of the input data needs to be improved to lower the frailty index value, thereby determining a key factor for improving the frailty index.

[0015] According to various embodiments of the present invention as described above, it is possible to predict the frailty indicators of the elderly in real time based on objective data, and to provide personalized lifestyle patterns that need improvement based on the predicted frailty indicators.

[0016] FIG. 1 is a schematic conceptual diagram for explaining the configuration of a real-time aging indicator prediction system based on bio-information and lifelog data according to an embodiment of the present invention.

[0017] FIG. 2 is a schematic block diagram illustrating the configuration of a real-time aging indicator prediction device based on bio-information and lifelog data according to an embodiment of the present invention.

[0018] FIG. 3 is a diagram illustrating the operation of a real-time aging indicator prediction device based on bio-information and lifelog data according to an embodiment of the present invention.

[0019] FIG. 4 is a flowchart illustrating a method for predicting real-time aging indicators based on bio-information and lifelog data according to an embodiment of the present invention, and,

[0020] FIG. 5 is a system diagram illustrating the operation of a real-time aging indicator prediction system based on bio-information and lifelog data according to one embodiment of the present invention.

[0021] Various embodiments of this document are described below with reference to the accompanying drawings. However, this is not intended to limit the technology described in this document to specific embodiments and should be understood to include various modifications, equivalents, and / or alternatives to the embodiments of this document. Similar reference numerals may be used for similar components in connection with the description of the drawings.

[0022] In this document, expressions such as 'have,' 'can have,' 'include,' or 'can include' refer to the existence of the relevant feature (e.g., components such as numerical values, functions, actions, or parts) and do not exclude the existence of additional features.

[0023] In this document, expressions such as 'A or B', 'at least one of A and / or B', or 'one or more of A and / or B' may include all possible combinations of items listed together. For example, 'A or B', 'at least one of A and B', or 'at least one of A or B' may refer to cases including (1) at least one A, (2) at least one B, or (3) both at least one A and at least one B. Expressions such as 'first', 'second', 'first', or 'second' used in this document may modify various components regardless of order and / or importance, and are used only to distinguish one component from another and do not limit said components.

[0024] As used in this document, the expression 'configured to' may be replaced, depending on the context, with, for example, 'suitable for,' 'having the capacity to,' 'designed to,' 'adapted to,' 'made to,' or 'capable of.' The term 'configured to' does not necessarily mean 'specifically designed to' in hardware. Instead, in some situations, the expression 'device configured to' may mean that the device is 'capable of' doing something together with other devices or components.

[0025] The terms used in this specification are for the purpose of describing embodiments and are not intended to limit or / or restrict the invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as "comprising" or "having" are intended to indicate the existence of the features, numbers, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, actions, components, parts, or combinations thereof.

[0026] In the embodiments, a 'module' or 'part' performs at least one function or operation and may be implemented in hardware or software, or a combination of hardware or software. Additionally, a plurality of 'modules' or a plurality of 'parts' may be integrated into at least one module and implemented by at least one processor, except for a 'module' or 'part' that needs to be implemented in specific hardware.

[0027] The present invention will be described in detail below using the attached drawings.

[0028] FIG. 1 is a schematic conceptual diagram for explaining the configuration of a real-time frailty indicator prediction system (1000) based on bio-information and lifelog data according to an embodiment of the present invention. Referring to FIG. 1, the real-time frailty indicator prediction system (1000) may include a real-time frailty indicator prediction device (100), a wearable device (200), and an external server (300).

[0029] A real-time frailty index prediction device (100) can predict a frailty index value by combining bio-information, life log data, clinical information, frailty index (FI) measured and recorded at a medical institution, medical treatment information, medication information, health checkup data, and data within the medical institution (diagnosis history, surgery details, etc.). Specific details will be described later.

[0030] The wearable device (200) can be worn on the user's body to collect the user's biometric information and lifelog data through sensors, etc. For example, the wearable device (200) may be a smartwatch, a smart ring, etc. And due to the processing limitations of the smartwatch, etc., the smartwatch, etc., may transmit the collected data to a smartphone, and the smartphone may then transmit it to the real-time aging indicator prediction device (100). The operation of the wearable device (200) will be described to include cases where multiple devices operate in cooperation in this way.

[0031] The wearable device (200) can display biometric information and lifelog data collected in real time through a display unit (not shown). Similarly, the wearable device (200) can display a frailty prediction value received from a real-time frailty indicator prediction device (100), a frailty state based on the frailty prediction value, and optimal recommended activities based on the frailty prediction value through a display unit (not shown).

[0032] An external server (300) refers to a server that holds the user's medical data, such as a medical institution or a National Health Insurance Service. The external server (300) can transmit the user's medical data necessary for frailty prediction upon a request from the real-time frailty index prediction device (100). The medical data held by the external server (300) may include clinical information, frailty index (FI) measured and recorded at a medical institution, medical treatment information, medication information, health checkup data, and data within the medical institution (diagnosis history, surgery details, etc.). In particular, clinical information may include the user's ID, gender, age, height, weight, medical history, nutritional status, grip strength, SPPB (Short Physical Performance Battery), BMI (Body Mass Index), etc.

[0033] FIG. 2 is a schematic block diagram for explaining the configuration of a real-time frailty indicator prediction device (100) according to an embodiment of the present invention. Referring to FIG. 2, the real-time frailty indicator prediction device (100) may include a communication unit (110), a database (120), and a processor (130). It goes without saying that other configurations, such as an input unit (not shown), may also be included.

[0034] The real-time frailty indicator prediction device (100) can be implemented in various forms. For example, the real-time frailty indicator prediction device (100) may be implemented as a separate electronic device to predict frailty indicator values ​​through communication with an external device or server and provide a function to recommend personalized optimal activities. As another example, the real-time frailty indicator prediction device (100) may be implemented in the form of a cloud server to provide functions such as predicting frailty indicator values, providing frailty status, and deriving key factors for improving frailty indicators in a SaaS format.

[0035] The communication unit (110) communicates with external devices or servers, etc. For example, the communication unit (110) can receive data such as biometric information, lifelog data, clinical information, frailty index (FI) measured and recorded at a medical institution, medical treatment information, medication information, health checkup data, and data within a medical institution (diagnosis history, surgery details, etc.) from a wearable device (200), an external server (300), or an external database. To do this, the communication unit (110) can use wired communication methods such as LAN, HDMI, etc., and wireless communication methods such as wireless LAN, NFC, IR communication, Zigbee communication, Bluetooth, etc.

[0036] The database (120) can store various modules, software, functions, artificial intelligence trained models, agents, data, etc. for operating the real-time frailty indicator prediction device (100). For example, the database (120) may store at least one artificial intelligence trained model for predicting frailty indicator values, and frailty indicator values ​​at the time of measurement per user may be stored in a time series.

[0037] The processor (130) can control the remaining configurations of the real-time aging indicator prediction device (100). For example, the processor (130) can control the communication unit (110) to receive biometric information and lifelog data from the wearable device (200). The processor (130) may be implemented as a single CPU to perform functions such as data processing, feature extraction, aging indicator prediction, and optimal activity recommendation, or it may be implemented as multiple processors and IPs or accelerators that perform specific functions. The operation of the specific processor (130) will be described below. For convenience when describing the operation of the processor (130), it should be noted that the operation may be divided into units such as the data management unit (131), feature extraction unit (132), aging indicator prediction unit (133), and optimal activity recommendation unit (134), as shown in FIG. 3.

[0038] The processor (130) can control the communication unit (110) to receive biometric information and lifelog data from the wearable device (200). The wearable device (200) can collect the user's biometric information and lifelog data in real time. For example, the biometric information may include blood pressure, blood sugar, heart rate, and ECG data. In addition, the lifelog data may include the user's step count, distance walked, and time walked.

[0039] The processor (130) can process the received biometric information and lifelog data and store them in the database (120). For example, the processor (130) can determine whether there are missing values ​​in each of the data constituting the biometric information and lifelog data. For each of the data determined to have missing values, the processor (130) can perform a data processing operation to predict and add the missing values. The processor (130) can use the data stored in the database (120) to predict the missing values. The database (120) may contain biometric information and lifelog data that has been previously received and stored. The processor (130) can predict and add missing values ​​in new data by utilizing the average value and standard deviation value for each of the data extracted from the biometric information and lifelog data that is previously stored. This operation can be expressed as an operation of the data management unit (131).

[0040] The processor (130) can correct the received biometric information using data received from an external server (300). The processor (130) can receive clinical information, a frailty index (FI) measured and recorded at a medical institution, medical treatment information, medication information, health checkup data, and data within the medical institution (diagnosis history, surgery details, etc.) from the external server (300). For example, the processor (130) can receive health checkup data from the external server (300). The received health checkup data may include information corresponding to the user's biometric information. The processor (130) can set data containing information corresponding to the user's biometric information as reference data. Then, the processor (130) can perform a processing operation to correct the collected biometric information using the set reference data. Through this, the processor (130) can solve the problem where the measured biometric information does not accurately reflect the user's biometric information due to temporary malfunction of the wearable device (200), the influence of the environment in which the user is located, etc. This operation can be expressed as an operation of the data management unit (131).

[0041] The processor (130) can generate input data by extracting characteristics from processed biometric information and lifelog data. First, in the case of biometric information that can be processed and stored in a time-series manner, the processor (130) can extract time-series characteristics from the processed biometric information. Time-series characteristics may include patterns of change over time, periods, outlier values, etc. For example, characteristics such as whether an abnormal pattern appears in an ECG can be extracted. This operation can be expressed as the operation of the characteristic extraction unit (132).

[0042] And the processor (130) can extract characteristics regarding whether there is a change, the amount of change, and the direction of change from the lifelog data. For example, it can extract change patterns and features regarding the amount of change in activity, the amount of change in walking speed, and the ratio of walking time from the number of steps, distance walked, and time walked that constitute the lifelog data. This operation can be expressed as the operation of the characteristic extraction unit (132).

[0043] Change in activity level may refer to the change in distance run or walked on a daily, weekly, or monthly basis. Change in walking speed refers to the change in whether walking speed has decreased or increased from the monthly average. Walking speed can be measured by dividing the distance walked by the time taken to walk, and if there is no previously stored average walking speed, the average walking speed of Korean elderly people may be applied. The average walking speed of Korean elderly people is 64 m per minute. The walking time ratio refers to the ratio of the time spent walking in daily life, excluding time spent sleeping, eating, etc., out of the total time spent moving. The processor (130) can extract characteristics regarding whether there is a change, the amount of change (degree of change), and the direction of change by comparing the change in activity level, the change in walking speed, and the walking time ratio with the daily, weekly, and monthly average values, respectively.

[0044] Next, the processor (130) can integrate the extracted features to generate input data to be input into the completed artificial intelligence model. Before explaining the subsequent operations, we will first explain the training phase of the completed artificial intelligence model.

[0045] In the present invention, a fully trained artificial intelligence model that performs the following operations may be used. First, the fully trained artificial intelligence model can output a predicted value for a frailty index. Second, the fully trained artificial intelligence model can output key factors for improving the frailty index and suggest optimal activities for improving the frailty index. Such a fully trained artificial intelligence model may be composed of a single artificial intelligence model or may be composed of multiple agents that perform each function.

[0046] The processor (130) can train an artificial intelligence model for predicting frailty using a Long Short-Term Memory (LSTM) algorithm. The artificial intelligence model trained using the LSTM algorithm receives the characteristics of time-series data as input, learns patterns over time using one or more LSTM layers, and can generate and output predicted values. The processor (130) can use biometric information and lifelog data obtained from a wearable device (200), as well as clinical information, medical treatment information, medication information, health checkup information, diagnosis history information, surgery details information, and frailty indicators (FI) tested at a medical institution received from an external server (300), as training data for training the artificial intelligence model. When frailty indicators (FI) tested at a medical institution are used as training data, the processor (130) can train the artificial intelligence model using a supervised learning method.

[0047] The processor (130) inputs the generated input data into a trained frailty prediction artificial intelligence model to receive outputs of frailty index prediction values ​​and key factors for improving frailty index. This operation can be expressed as the operation of the frailty index prediction unit (133). The frailty index value predicted by the trained artificial intelligence model represents the degree of frailty of each user in the current situation. As these prediction values ​​accumulate, changes in the frailty index value can be known, which means that it is possible to know whether the user's condition is improving, worsening, or unchanged.

[0048] The predicted frailty index value is represented as a number between 0 and 1. If the predicted frailty index value is 0 or greater and less than 0.12, it corresponds to a normal state. If the predicted frailty index value is 0.12 or greater and less than 0.24, it may be classified as a pre-frailty state; if it is 0.24 or greater and less than 0.36, it may be classified as a moderate frailty state; and if it is 0.36 or greater, it may be classified as a severe frailty state. The processor (130) can control the communication unit (110) to transmit the predicted frailty index value and the frailty state classification value (normal, pre-frailty, moderate frailty, severe frailty) to the wearable device (200). The wearable device (200) configures a screen using the received frailty index value or frailty state classification value, and displays the configured screen through a display unit (not shown) so that the user can recognize their current frailty state and changes.

[0049] The processor (130) can derive key factors for improving the frailty index. The processor (130) can derive optimal activities for improving the frailty index using the derived key factors. This operation can be expressed as the operation of the optimal activity recommendation unit (134). Specifically, the processor (130) can determine which data among the input data needs to be improved in order to lower the frailty index value. Through this, the processor (130) can determine key factors for improving the frailty index. The processor (130) can derive key factors based on (1) frailty index values ​​predicted by an artificial intelligence learning completed model, (2) medical information, medication information, and health checkup information received from an external server (300), (3) biometric information received from a wearable device (200) (e.g., only biometric information for a specific period such as the last 1 month and 3 months may be used), and (4) changes in activity amount, changes in walking speed, and walking time ratios extracted by processing lifelog data. The processor (130) can suggest steps, walking distance, walking speed, and walking time for each user through the change in the aging index that changes in response to each factor. The information including the suggested steps, walking distance, walking speed, and walking time may correspond to the recommended optimal information.

[0050] The processor (130) can transmit information about the derived optimal activity to the wearable device (200) via the communication unit (110). At this time, the information about the optimal activity may also include information on how many points the user's aging index can be lowered by if the recommended optimal activity is performed. The wearable device (200) can display the received optimal activity information to encourage the user to perform the corresponding action. The wearable device (200) can determine how much the user has performed the activity corresponding to the displayed optimal activity information through a built-in sensor. And the wearable device (200) can display an estimated value of how much the aging index will be in response to the user's activity performance through a display unit (not shown).

[0051] According to various embodiments of the present invention as described above, there is an effect that frailty indicators, which previously had to be measured by visiting a medical institution and conducting a survey, can be obtained simply by wearing a wearable device. Furthermore, there is an effect that frailty indicators can be determined based on objective data rather than being determined by subjective responses. Through this, personalized optimal activities can be suggested, and elderly people who follow the suggested optimal activities can improve their degree of frailty.

[0052] FIG. 4 is a flowchart illustrating a method for predicting real-time frailty indicators based on biometric information and lifelog data according to an embodiment of the present invention. Referring to FIG. 4, a real-time frailty indicator prediction system (1000) can collect and process biometric information and lifelog data (S410). Specifically, the real-time frailty indicator prediction system (1000) can perform processing to add missing values ​​to the collected biometric information and lifelog data using the mean and standard deviation extracted from the previously stored biometric information and lifelog data. In addition, the real-time frailty indicator prediction system (1000) can set data containing information corresponding to biometric information, such as health checkup data, as reference data, and perform processing to correct the collected biometric information using the set reference data.

[0053] Next, the real-time frailty indicator prediction system (1000) can generate input data by extracting characteristics from processed biometric information and lifelog data (S420). Specifically, the real-time frailty indicator prediction system (1000) can extract time-series characteristics from processed biometric information. And the real-time frailty indicator prediction system (1000) can extract characteristics regarding whether there is a change, the amount of change, and the direction of change from processed lifelog data. By integrating these extracted characteristics, the real-time frailty indicator prediction system (1000) can generate input data to be fed into a trained artificial intelligence model.

[0054] Next, the real-time frailty index prediction system (1000) can input the generated input data into a frailty prediction artificial intelligence model to receive outputs of a frailty index prediction value and a key factor for improving the frailty index (S430). Specifically, the real-time frailty prediction system (1000) can input the input data into an artificial intelligence model trained using an LSTM algorithm to output a frailty index prediction value. Then, the real-time frailty prediction system (1000) can determine which of the input data needs to be improved to lower the frailty index value and determine a key factor for improving the frailty index.

[0055] FIG. 5 is a system diagram for explaining the operation of a real-time aging indicator prediction system (1000) based on bio-information and lifelog data according to an embodiment of the present invention. Overlapping operations will be described briefly, and the operation between the devices (100, 200, 300) constituting the system (1000) will be described mainly.

[0056] Referring to FIG. 5, the wearable device (200) can collect biometric information and lifelog data (S505). The wearable device (200) can then transmit the collected biometric information and lifelog data to the aging indicator prediction device (100) (S510).

[0057] At the same time or sequentially, the external server (300) can collect health checkup data, medication information, and frailty indicator (FI) information obtained from a medical institution (S515). Then, the external server (300) can transmit the collected health checkup data, medication information, and frailty indicator (FI) information to the frailty indicator prediction device (100) (S520). At this time, only health checkup data, medication information, and frailty indicator (FI) information are listed as examples, and data such as clinical information, medical treatment information, diagnosis history, and surgery details are not excluded.

[0058] A frailty indicator prediction device (100), having received data from a wearable device (200) and an external server (300), first corrects biometric information and can generate processed data using lifelog data and corrected biometric information (S525). Subsequently, the frailty indicator prediction device (100) can extract features from the processed data to generate input data to be input into a learned artificial intelligence (S530). The frailty indicator prediction device (100) can input the generated input data into the artificial intelligence model to output a frailty indicator prediction value and an indicator improvement factor (S535). The frailty indicator prediction device (100) can derive activities for each user to optimally improve their frailty indicator using the outputted frailty indicator improvement factor (S540). The derived customized optimal activity information is then transmitted from the frailty indicator prediction device (100) to the wearable device (200) (S545). The wearable device (200) can configure a screen using the received customized optimal activity information and display the optimal activity to the user (S550).

[0059] The methods described above 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 devices may be configured to operate as one or more software modules to perform the operation of the present invention, and vice versa.

[0060] As described above, although the present disclosure has been explained by limited embodiments and drawings, the present disclosure is not limited to the above embodiments, and various modifications and variations are possible from this description by those skilled in the art to which the present disclosure belongs. Therefore, the scope of the present disclosure should not be limited to the described embodiments, but should be defined by the claims set forth below as well as equivalents thereof.

Claims

1. In a method for predicting real-time frailty indicators based on biometric information and lifelog data, A step of collecting and processing biometric information and lifelog data; A step of generating input data by extracting characteristics from the above-mentioned processed biometric information and lifelog data; and A method for predicting a frailty indicator, comprising the step of inputting the above-mentioned generated input data into a frailty prediction artificial intelligence model to receive outputs of a frailty indicator prediction value and a key factor for improving the frailty indicator.

2. In Paragraph 1, The above-mentioned step of collecting and processing is, A method for predicting aging indicators by processing the collected biometric information and lifelog data by adding missing values ​​using the mean and standard deviation extracted from the previously stored biometric information and lifelog data.

3. In Paragraph 1, The above-mentioned step of collecting and processing is, A method for predicting aging indicators by setting health checkup data received from an external server as reference data and processing the collected biometric information using the set reference data.

4. In Paragraph 1, The step of generating the above input data is, A method for predicting aging indicators that extracts time-series characteristics from the processed bio-information, extracts characteristics regarding whether there is a change, the amount of change, and the direction of change from the processed life-log data, and integrates the extracted characteristics to generate the input data.

5. In Paragraph 1, The above output receiving step is, The above input data is input into an artificial intelligence model trained using an LSTM algorithm to output a predicted value for the frailty index, and A method for predicting a frailty index that determines which of the input data needs to be improved to lower the above frailty index value, and determines key factors for improving the frailty index.

6. In a device for predicting real-time aging indicators based on biometric information and lifelog data, A communication unit that communicates with an external device; A database that stores data; and A processor that controls the above-mentioned aging indicator prediction device; comprising The above processor is, A frailty indicator prediction device that controls a communication unit to receive biometric information and lifelog data collected from a wearable device, processes the received biometric information and lifelog data, extracts features from the processed biometric information and lifelog data to generate input data, inputs the generated input data into a frailty prediction artificial intelligence model to output a frailty indicator prediction value and a key factor for improving the frailty indicator, and stores them in the database.

7. In Paragraph 6, The above processor is, A aging indicator prediction device that performs processing to add missing values ​​to the above-mentioned collected biometric information and lifelog data using the mean and standard deviation extracted from the previously stored biometric information and lifelog data.

8. In Paragraph 6, The above processor is, A aging indicator prediction device that controls the communication unit to receive health checkup data from an external server, sets the received health checkup data as reference data, and performs processing to correct the collected biometric information using the set reference data.

9. In Paragraph 6, The above processor is, A aging indicator prediction device that extracts time-series characteristics from the processed bio-information, extracts characteristics regarding whether there is a change, the amount of change, and the direction of change from the processed life-log data, and integrates the extracted characteristics to generate the input data.

10. In Paragraph 6, The above processor is, A frailty index prediction device that inputs the above input data into an artificial intelligence model trained using an LSTM algorithm to output a frailty index prediction value, and determines which of the input data needs to be improved to lower the frailty index value, thereby determining a key factor for improving the frailty index.

Images