Method and device for predicting health index and disease of companion animal through artificial intelligence

Abstract

Embodiments present a method and device for providing a service for predicting the health index and disease of a companion animal through artificial intelligence. A device according to one embodiment may comprise: a companion animal information reception unit for receiving basic information about a companion animal and inspection information about the companion animal; a health index determination unit for determining a health index for each of a plurality of items of the companion animal by using a first machine learning model on the basis of the basic information about the companion animal and the inspection information about the companion animal; a disease prediction unit for determining prediction information about diseases of the companion animal by using a second machine learning model on the basis of the basic information about the companion animal and the inspection information about the companion animal; and a report generation unit for automatically generating a result report including the health index for each of the plurality of items of the companion animal and the prediction information about the diseases of the companion animal.

Description

Method and device for predicting the health index and disease of companion animals using artificial intelligence

[0001] Embodiments of the present disclosure relate to a technology for predicting the health index and disease of a companion animal, and to a technology for predicting the health index and disease of a companion animal using artificial intelligence.

[0002] Meanwhile, due to social changes such as low birth rates, aging population, and increase in single-person households, an era has arrived in which 6.04 million households, or about a quarter of the domestic population, care for pets. According to one statistical survey, 85.6% of households raising pets agreed with the statement that "pets are family," and recently, terms such as "PetFam" and "PetMily" have emerged, referring to households that do not spare investment in their pets. As such, the number of households adopting pets continues to increase, and this can be attributed mainly to the continued decrease in the number of people per household and the increased emotional stability and happiness within the household due to pets.

[0003] Comparing the past decade with the past decade since 2000, pet adoptions have significantly increased in the past decade, while pets over 10 years of age still represent a relatively small proportion. This suggests that the number of aging companion animals will increase in the future, further expanding their market potential. The market for preventative diagnostics for healthy living is expected to experience rapid growth in the future.

[0004] In particular, the lack of disease prevention and diagnosis technology in animal hospitals is a great stressor for pet owners who prioritize the welfare of their pets, who are like family members.

[0005] In addition, as the proportion of pets over 10 years old increases, the need for health management due to the aging of pets becomes necessary, and as the proportion of expenditure on pet medical services (disease / preventive treatment costs) accounts for 20% of the cost of raising pets, there are cases where pet ownership is abandoned due to the financial burden, and the need for prevention is emerging.

[0006] However, there is no health checkup service that can monitor the health of pets before they become ill, and medical services for various diagnoses after an illness occurs (MRI, X-ray, etc.) are not covered by insurance and are therefore more expensive than those for humans.

[0007] Accordingly, a method and device for predicting the health index and disease of companion animals using artificial intelligence learned based on the companion animal's medical data and hair data are needed.

[0008] Embodiments of the present disclosure can provide a method and device for predicting the health index and disease of a companion animal through artificial intelligence.

[0009] The technical tasks to be achieved in the embodiments are not limited to those mentioned above, and other technical tasks not mentioned can be considered by a person having ordinary skill in the art from the various embodiments described below.

[0010] A device for providing a service for predicting a health index and disease of a companion animal through artificial intelligence according to one embodiment may include a companion animal information receiving unit that receives basic information and examination information of the companion animal, a health index determining unit that determines a plurality of health indices of the companion animal by using a first machine learning model based on the basic information and the examination information of the companion animal, a disease prediction unit that determines prediction information on a disease of the companion animal by using a second machine learning model based on the basic information and the examination information of the companion animal, and a report generating unit that automatically generates a result report including the plurality of health indices of the companion animal by items and the prediction information on the disease of the companion animal. For example, the result report may be transmitted to a user terminal. For example, among a plurality of first examination variables related to hair analysis results and a plurality of second examination variables related to blood analysis results, an examination variable having a correlation score for a health index greater than or equal to a preset score is determined as a first influence variable for each of the plurality of items, and the first influence variable may be used as learning data of the first machine learning model. For example, among the plurality of first test variables and the plurality of second test variables, a test variable whose correlation score for a disease is greater than or equal to a preset score is determined as a second influential variable for each disease, and the second influential variable can be used as learning data for the second machine learning model.

[0011] According to embodiments, a customized report predicting the disease and health index of a pet can be provided to a user terminal through artificial intelligence learned based on nutritional mineral and harmful heavy metal data analyzed using a precision analysis device (such as ICP-MS) using a small amount of pet fur as a sample and medical data obtained from an animal hospital.

[0012] The effects that can be obtained from the examples are not limited to the effects mentioned above, and other effects that are not mentioned can be clearly derived and understood by a person having ordinary skill in the art based on the detailed description below.

[0013] The accompanying drawings, which are included as part of the detailed description to aid understanding of the embodiments, provide various embodiments and, together with the detailed description, describe technical features of the various embodiments.

[0014] FIG. 1 is a diagram showing the configuration of an electronic device according to one embodiment.

[0015] Figure 2 is a diagram showing the configuration of a program according to one embodiment.

[0016] FIG. 3 is a diagram of a system for providing a service for predicting the health index and disease of a companion animal through artificial intelligence according to one embodiment.

[0017] Fig. 4 is a block diagram exemplarily showing a functional module of a service providing device according to Fig. 3.

[0018] Figure 5 is a flowchart showing the process by which the health index determination unit according to Figure 4 determines the health index of a companion animal for each of multiple items.

[0019] FIG. 6 is an example of an algorithm of the first machine learning model used in the health index determination unit according to FIG. 4.

[0020] Fig. 7 is an example of the learning result of the algorithm of the first machine learning model used in the health index determination unit according to Fig. 4.

[0021] Figure 8 is a flowchart showing the process by which the disease prediction unit according to Figure 4 determines prediction information for a companion animal's disease.

[0022] FIG. 9 is an example of an algorithm of a second machine learning model used in the disease prediction unit according to FIG. 4.

[0023] Figure 10 is an example of the learning results of the algorithm of the second machine learning model used in the disease prediction unit according to Figure 4.

[0024] Figure 11 is a flowchart showing the process of determining a care method for a companion animal using a neural network by the care method recommendation unit according to Figure 4.

[0025] Figure 12 shows an example of a result report generated by the report generation unit according to Figure 4.

[0026] Fig. 13 is a block diagram showing the configuration of a server according to one embodiment.

[0027] Figure 14 is another example of the learning results of the algorithm of the second machine learning model used in the disease prediction unit according to Figure 4.

[0028] The following embodiments combine components and features of the embodiments in a predetermined form. Each component or feature may be considered optional unless explicitly stated otherwise. Each component or feature may be implemented without being combined with other components or features. Furthermore, various embodiments may be formed by combining some components and / or features. The order of operations described in various embodiments may be changed. Some components or features of one embodiment may be included in another embodiment or may be replaced with corresponding components or features of another embodiment.

[0029] In the description of the drawings, procedures or steps that may obscure the gist of various embodiments are not described, and procedures or steps that can be understood by a person with ordinary skill in the art are also not described.

[0030] Throughout the specification, when a part is said to "comprising" (or including) a certain component, this does not mean that other components are excluded, but rather that other components can be included, unless specifically stated otherwise. In addition, terms such as "...part," "...unit," and "module" described in the specification mean a unit that processes at least one function or operation, which may be implemented by hardware, software, or a combination of hardware and software. In addition, the words "a" or "an," "one," "the," and similar related words may be used in the singular and plural sense in the context of describing various embodiments (especially in the context of the claims below) unless otherwise indicated herein or clearly contradicted by context.

[0031] Hereinafter, embodiments according to various embodiments will be described in detail with reference to the attached drawings. The detailed description disclosed below, together with the attached drawings, is intended to explain exemplary embodiments of various embodiments and is not intended to represent the only embodiment.

[0032] Additionally, specific terms used in various embodiments are provided to aid understanding of the various embodiments, and the use of such specific terms may be changed in other forms without departing from the technical spirit of the various embodiments.

[0033] FIG. 1 is a diagram showing the configuration of an electronic device according to one embodiment.

[0034] FIG. 1 is a block diagram of an electronic device (101) within a network environment (100) according to various embodiments. Referring to FIG. 1, in the network environment (100), the electronic device (101) may communicate with the electronic device (102) via a first network (198) (e.g., a short-range wireless communication network), or may communicate with at least one of the electronic device (104) or the server (108) via a second network (199) (e.g., a long-range wireless communication network). In one embodiment, the electronic device (101) may communicate with the electronic device (104) via the server (108). According to one embodiment, the electronic device (101) may include a processor (120), a memory (130), an input module (150), an audio output module (155), a display module (160), an audio module (170), a sensor module (176), an interface (177), a connection terminal (178), a haptic module (179), a camera module (180), a power management module (188), a battery (189), a communication module (190), a subscriber identification module (196), or an antenna module (197). In some embodiments, the electronic device (101) may omit at least one of these components (e.g., the connection terminal (178)), or may have one or more other components added. In some embodiments, some of these components (e.g., the sensor module (176), the camera module (180), or the antenna module (197)) may be integrated into one component (e.g., the display module (160)). The electronic device (101) may also be referred to as a client, terminal, or peer.

[0035] The processor (120) may, for example, execute software (e.g., a program (140)) to control at least one other component (e.g., a hardware or software component) of the electronic device (101) connected to the processor (120) and perform various data processing or operations. According to one embodiment, as at least a part of the data processing or operations, the processor (120) may store commands or data received from other components (e.g., a sensor module (176) or a communication module (190)) in a volatile memory (132), process the commands or data stored in the volatile memory (132), and store result data in a non-volatile memory (134). According to one embodiment, the processor (120) may include a main processor (121) (e.g., a central processing unit or an application processor) or an auxiliary processor (123) (e.g., a graphics processing unit, a neural processing unit (NPU), an image signal processor, a sensor hub processor, or a communication processor) that can operate independently or together with the main processor (121). For example, when the electronic device (101) includes the main processor (121) and the auxiliary processor (123), the auxiliary processor (123) may be configured to use less power than the main processor (121) or to be specialized for a given function. The auxiliary processor (123) may be implemented separately from the main processor (121) or as a part thereof.

[0036] The auxiliary processor (123) may control at least a portion of functions or states associated with at least one component (e.g., a display module (160), a sensor module (176), or a communication module (190)) of the electronic device (101), for example, on behalf of the main processor (121) while the main processor (121) is in an inactive (e.g., sleep) state, or together with the main processor (121) while the main processor (121) is in an active (e.g., application execution) state. In one embodiment, the auxiliary processor (123) (e.g., an image signal processor or a communication processor) may be implemented as a part of another functionally related component (e.g., a camera module (180) or a communication module (190)). In one embodiment, the auxiliary processor (123) (e.g., a neural network processing unit) may include a hardware structure specialized for processing artificial intelligence models.

[0037] An artificial intelligence model can be generated through machine learning. Such learning can be performed, for example, in the electronic device (101) on which the artificial intelligence model is executed, or can be performed through a separate server (e.g., server (108)). The learning algorithm can include, for example, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but is not limited to the examples described above. The artificial intelligence model can include a plurality of artificial neural network layers. The artificial neural network can be one of a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent deep neural network (BRDNN), a deep Q-network, or a combination of two or more of the above, but is not limited to the examples described above. An artificial intelligence model may additionally or alternatively include a software structure in addition to a hardware structure.

[0038] The memory (130) can store various data used by at least one component (e.g., processor (120) or sensor module (176)) of the electronic device (101). The data can include, for example, software (e.g., program (140)) and input data or output data for commands related thereto. The memory (130) can include volatile memory (132) or non-volatile memory (134).

[0039] The program (140) may be stored as software in the memory (130) and may include, for example, an operating system (142), middleware (144), or an application (146).

[0040] The input module (150) can receive commands or data to be used in a component of the electronic device (101) (e.g., a processor (120)) from an external source (e.g., a user) of the electronic device (101). The input module (150) can include, for example, a microphone, a mouse, a keyboard, a key (e.g., a button), or a digital pen (e.g., a stylus pen).

[0041] The audio output module (155) can output audio signals to the outside of the electronic device (101). The audio output module (155) can include, for example, a speaker or a receiver. The speaker can be used for general purposes, such as multimedia playback or recording playback. The receiver can be used to receive incoming calls. In one embodiment, the receiver can be implemented separately from the speaker or as part of the speaker.

[0042] The display module (160) can visually provide information to an external party (e.g., a user) of the electronic device (101). The display module (160) may include, for example, a display, a holographic device, or a projector and a control circuit for controlling the device. According to one embodiment, the display module (160) may include a touch sensor configured to detect a touch, or a pressure sensor configured to measure the intensity of a force generated by the touch.

[0043] The audio module (170) can convert sound into an electrical signal, or vice versa, convert an electrical signal into sound. According to one embodiment, the audio module (170) can acquire sound through the input module (150), output sound through the sound output module (155), or an external electronic device (e.g., electronic device (102)) (e.g., speaker or headphone) directly or wirelessly connected to the electronic device (101).

[0044] The sensor module (176) can detect the operating status (e.g., power or temperature) of the electronic device (101) or the external environmental status (e.g., user status) and generate an electrical signal or data value corresponding to the detected status. According to one embodiment, the sensor module (176) can include, for example, a gesture sensor, a gyro sensor, a barometric pressure sensor, a magnetic sensor, an acceleration sensor, a grip sensor, a proximity sensor, a color sensor, an IR (infrared) sensor, a biometric sensor, a temperature sensor, a humidity sensor, or an illuminance sensor.

[0045] The interface (177) may support one or more designated protocols that may be used to directly or wirelessly connect the electronic device (101) with an external electronic device (e.g., the electronic device (102)). In one embodiment, the interface (177) may include, for example, a high definition multimedia interface (HDMI), a universal serial bus (USB) interface, an SD card interface, or an audio interface.

[0046] The connection terminal (178) may include a connector through which the electronic device (101) may be physically connected to an external electronic device (e.g., electronic device (102)). According to one embodiment, the connection terminal (178) may include, for example, an HDMI connector, a USB connector, an SD card connector, or an audio connector (e.g., a headphone connector).

[0047] The haptic module (179) can convert electrical signals into mechanical stimuli (e.g., vibration or movement) or electrical stimuli that a user can perceive through tactile or kinesthetic sensations. According to one embodiment, the haptic module (179) can include, for example, a motor, a piezoelectric element, or an electrical stimulation device.

[0048] The camera module (180) can capture still images and videos. According to one embodiment, the camera module (180) may include one or more lenses, image sensors, image signal processors, or flashes.

[0049] The power management module (188) can manage power supplied to the electronic device (101). According to one embodiment, the power management module (188) can be implemented as, for example, at least a part of a power management integrated circuit (PMIC).

[0050] A battery (189) may power at least one component of the electronic device (101). In one embodiment, the battery (189) may include, for example, a non-rechargeable primary battery, a rechargeable secondary battery, or a fuel cell.

[0051] The communication module (190) may support the establishment of a direct (e.g., wired) communication channel or a wireless communication channel between the electronic device (101) and an external electronic device (e.g., electronic device (102), electronic device (104), or server (108)), and the performance of communication through the established communication channel. The communication module (190) may operate independently from the processor (120) (e.g., application processor) and may include one or more communication processors that support direct (e.g., wired) communication or wireless communication. According to one embodiment, the communication module (190) may include a wireless communication module (192) (e.g., a cellular communication module, a short-range wireless communication module, or a global navigation satellite system (GNSS) communication module) or a wired communication module (194) (e.g., a local area network (LAN) communication module, or a power line communication module). Among these communication modules, the corresponding communication module can communicate with an external electronic device (104) via a first network (198) (e.g., a short-range communication network such as Bluetooth, wireless fidelity (WiFi) direct, or infrared data association (IrDA)) or a second network (199) (e.g., a long-range communication network such as a legacy cellular network, a 5G network, a next-generation communication network, the Internet, or a computer network (e.g., a LAN or WAN)). These various types of communication modules can be integrated into a single component (e.g., a single chip) or implemented as multiple separate components (e.g., multiple chips). The wireless communication module (192) can verify or authenticate the electronic device (101) within a communication network such as the first network (198) or the second network (199) by using subscriber information (e.g., an international mobile subscriber identity (IMSI)) stored in the subscriber identification module (196).

[0052] The wireless communication module (192) can support 5G networks and next-generation communication technologies following the 4G network, such as NR access technology (new radio access technology). The NR access technology can support high-speed transmission of high-capacity data (eMBB (enhanced mobile broadband)), minimization of terminal power and connection of multiple terminals (mMTC (massive machine type communications)), or high reliability and low latency (URLLC (ultra-reliable and low-latency communications)). The wireless communication module (192) can support, for example, a high-frequency band (e.g., mmWave band) to achieve a high data transmission rate. The wireless communication module (192) can support various technologies for securing performance in a high-frequency band, such as beamforming, massive multiple-input and multiple-output (MIMO), full dimensional MIMO (FD-MIMO), array antenna, analog beam-forming, or large scale antenna. The wireless communication module (192) can support various requirements specified in the electronic device (101), an external electronic device (e.g., the electronic device (104)), or a network system (e.g., the second network (199)). According to one embodiment, the wireless communication module (192) can support a peak data rate (e.g., 20 Gbps or more) for eMBB realization, a loss coverage (e.g., 164 dB or less) for mMTC realization, or a U-plane latency (e.g., 0.5 ms or less for downlink (DL) and uplink (UL), or 1 ms or less for round trip) for URLLC realization.

[0053] The antenna module (197) can transmit or receive signals or power to or from an external device (e.g., an external electronic device). In one embodiment, the antenna module (197) may include an antenna including a radiator formed of a conductor or a conductive pattern formed on a substrate (e.g., a PCB). In one embodiment, the antenna module (197) may include a plurality of antennas (e.g., an array antenna). In this case, at least one antenna suitable for a communication method used in a communication network, such as the first network (198) or the second network (199), may be selected from the plurality of antennas, for example, by the communication module (190). A signal or power may be transmitted or received between the communication module (190) and an external electronic device via the at least one selected antenna. In some embodiments, in addition to the radiator, another component (e.g., a radio frequency integrated circuit (RFIC)) may be additionally formed as a part of the antenna module (197).

[0054] According to various embodiments, the antenna module (197) may form a mmWave antenna module. In one embodiment, the mmWave antenna module may include a printed circuit board, an RFIC disposed on or adjacent a first side (e.g., a bottom side) of the printed circuit board and capable of supporting a designated high-frequency band (e.g., a mmWave band), and a plurality of antennas (e.g., an array antenna) disposed on or adjacent a second side (e.g., a top side or a side side) of the printed circuit board and capable of transmitting or receiving signals in the designated high-frequency band.

[0055] At least some of the above components can be interconnected and exchange signals (e.g., commands or data) with each other via a communication method between peripheral devices (e.g., a bus, GPIO (general purpose input and output), SPI (serial peripheral interface), or MIPI (mobile industry processor interface)).

[0056] According to one embodiment, commands or data may be transmitted or received between the electronic device (101) and an external electronic device (104) via a server (108) connected to a second network (199). Each of the external electronic devices (102 or 104) may be the same or a different type of device as the electronic device (101). According to one embodiment, all or part of the operations executed in the electronic device (101) may be executed in one or more of the external electronic devices (102, 104, or 108). For example, when the electronic device (101) is to perform a certain function or service automatically or in response to a request from a user or another device, the electronic device (101) may, instead of or in addition to executing the function or service itself, request one or more external electronic devices to perform the function or at least a part of the service. One or more external electronic devices that receive the request may execute at least a portion of the requested function or service, or an additional function or service related to the request, and transmit the result of the execution to the electronic device (101). The electronic device (101) may process the result as is or additionally and provide it as at least a portion of a response to the request. For this purpose, cloud computing, distributed computing, mobile edge computing (MEC), or client-server computing technology may be used, for example. The electronic device (101) may provide an ultra-low latency service by using distributed computing or mobile edge computing, for example. In another embodiment, the external electronic device (104) may include an Internet of Things (IoT) device. The server (108) may be an intelligent server utilizing machine learning and / or a neural network. According to one embodiment, the external electronic device (104) or the server (108) may be included in the second network (199).The electronic device (101) can be applied to intelligent services (e.g., smart home, smart city, smart car, or healthcare) based on 5G communication technology and IoT-related technology.

[0057] The server (108) is connected to an electronic device (101) and can provide services to the connected electronic device (101). In addition, the server (108) can process a membership registration process, store and manage various information of users who have registered as members, and provide various purchase and payment functions related to the service. In addition, the server (108) can share in real time the execution data of service applications running on each of a plurality of electronic devices (101) so that services can be shared among users. This server (108) may have the same hardware configuration as a typical web server or service server. However, in terms of software, it may include program modules that are implemented in any language such as C, C++, Java, Python, Golang, or Kotlin and perform various functions. In addition, the server (108) generally refers to a computer system and computer software (server program) installed therefor that is connected to an unspecified number of clients and / or other servers through an open computer network such as the Internet, and that receives a request to perform a task from a client or other server and provides the result of the task accordingly. In addition, the server (108) should be understood as a broad concept that includes, in addition to the above-mentioned server program, a series of application programs running on the server (108) and, in some cases, various databases (DB: Database, hereinafter referred to as “DB”) built internally or externally. Accordingly, the server (108) classifies member registration information and various information and data about the game and stores and manages them in the DB, and this DB can be implemented internally or externally to the server (108).In addition, the server (108) can be implemented using a variety of server programs provided according to operating systems such as Windows, Linux, UNIX, and Macintosh on general server hardware, and representative examples include IIS (Internet Information Server) used in a Windows environment and CERN, NCSA, APPACH, TOMCAT, etc. used in a UNIX environment, which can implement a web service. In addition, the server (108) can also be linked with an authentication system and a payment system for user authentication of a service or purchase payment related to a service.

[0058] The first network (198) and the second network (199) refer to a connection structure that enables information exchange between each node, such as terminals and servers, or a network that connects the server (108) and electronic devices (101, 104). The first network (198) and the second network (199) include, but are not limited to, the Internet, a Local Area Network (LAN), a Wireless Local Area Network (Wireless LAN), a Wide Area Network (WAN), a Personal Area Network (PAN), 3G, 4G, LTE, 5G, Wi-Fi, etc. The first network (198) and the second network (199) may be closed first networks (198) and second networks (199) such as LANs and WANs, but are preferably open such as the Internet. The Internet refers to a worldwide open computer network (198) and second network (199) structure that provides protocols such as TCP / IP protocol, TCP, UDP (user datagram protocol), and various services existing at their upper layers, namely HTTP (HyperText Transfer Protocol), Telnet, FTP (File Transfer Protocol), DNS (Domain Name System), SMTP (Simple Mail Transfer Protocol), SNMP (Simple Network Management Protocol), NFS (Network File Service), and NIS (Network Information Service).

[0059] A database can have a general data structure implemented in the storage space (hard disk or memory) of a computer system using a database management program (DBMS). The database can have a data storage form that allows free searching (extracting), deleting, editing, adding, etc. of data. The database can be implemented to suit the purpose of one embodiment of the present disclosure using a relational database management system (RDBMS) such as Oracle, Informix, Sybase, or DB2, an object-oriented database management system (OODBMS) such as Gemston, Orion, or O2, and an XML native database such as Excelon, Tamino, or Sekaiju, and can have appropriate fields or elements to achieve its own function.

[0060] Figure 2 is a diagram showing the configuration of a program according to one embodiment.

[0061] FIG. 2 is a block diagram (200) illustrating a program (140) according to various embodiments. According to one embodiment, the program (140) may include an operating system (142), middleware (144), or an application (146) executable on the operating system (142) for controlling one or more resources of the electronic device (101). The operating system (142) may include, for example, Android™, iOS™, Windows™, Symbian™, Tizen™, or Bada™. At least some of the programs (140) may be preloaded onto the electronic device (101), for example, during manufacturing, or may be downloaded or updated from an external electronic device (e.g., the electronic device (102 or 104), or a server (108)) when used by a user. All or part of the program (140) may include a neural network.

[0062] The operating system (142) may control the management (e.g., allocation or retrieval) of one or more system resources (e.g., processes, memory, or power) of the electronic device (101). The operating system (142) may additionally or alternatively include one or more driver programs for driving other hardware devices of the electronic device (101), for example, an input module (150), an audio output module (155), a display module (160), an audio module (170), a sensor module (176), an interface (177), a haptic module (179), a camera module (180), a power management module (188), a battery (189), a communication module (190), a subscriber identification module (196), or an antenna module (197).

[0063] Middleware (144) can provide various functions to the application (146) so that functions or information provided from one or more resources of the electronic device (101) can be used by the application (146). Middleware (144) can include, for example, an application manager (201), a window manager (203), a multimedia manager (205), a resource manager (207), a power manager (209), a database manager (211), a package manager (213), a connectivity manager (215), a notification manager (217), a location manager (219), a graphics manager (221), a security manager (223), a call manager (225), or a voice recognition manager (227).

[0064] The application manager (201) can manage, for example, the life cycle of the application (146). The window manager (203) can manage, for example, one or more GUI resources used on the screen. The multimedia manager (205) can, for example, identify one or more formats required for playing media files, and perform encoding or decoding of a corresponding media file among the media files using a codec suitable for the corresponding format selected among the formats. The resource manager (207) can manage, for example, the source code of the application (146) or the memory space of the memory (130). The power manager (209) can manage, for example, the capacity, temperature, or power of the battery (189), and determine or provide related information necessary for the operation of the electronic device (101) using the corresponding information. According to one embodiment, the power manager (209) can be linked with the basic input / output system (BIOS) (not shown) of the electronic device (101).

[0065] The database manager (211) can, for example, create, search, or modify a database to be used by the application (146). The package manager (213) can, for example, manage the installation or update of an application distributed in the form of a package file. The connectivity manager (215) can, for example, manage a wireless connection or direct connection between the electronic device (101) and an external electronic device. The notification manager (217) can, for example, provide a function for notifying a user of the occurrence of a specified event (e.g., an incoming call, a message, or an alarm). The location manager (219) can, for example, manage location information of the electronic device (101). The graphics manager (221) can, for example, manage one or more graphic effects to be provided to the user or a user interface related thereto.

[0066] The security manager (223) may provide, for example, system security or user authentication. The telephony manager (225) may manage, for example, a voice call function or a video call function provided by the electronic device (101). The voice recognition manager (227) may, for example, transmit the user's voice data to the server (108) and receive, from the server (108), a command corresponding to a function to be performed in the electronic device (101) based at least in part on the voice data, or text data converted based at least in part on the voice data. In one embodiment, the middleware (244) may dynamically delete some existing components or add new components. In one embodiment, at least a portion of the middleware (144) may be included as a part of the operating system (142) or implemented as separate software different from the operating system (142).

[0067] The application (146) may include, for example, a home (251), a dialer (253), an SMS / MMS (255), an instant message (IM) (257), a browser (259), a camera (261), an alarm (263), a contact (265), a voice recognition (267), an email (269), a calendar (271), a media player (273), an album (275), a watch (277), a health (279) (e.g., measuring biometric information such as the amount of exercise or blood sugar), or an environmental information (281) (e.g., measuring barometric pressure, humidity, or temperature information) application. According to one embodiment, the application (146) may further include an information exchange application (not shown) that can support information exchange between the electronic device (101) and an external electronic device. The information exchange application may include, for example, a notification relay application configured to transmit designated information (e.g., a call, a message, or an alarm) to an external electronic device, or a device management application configured to manage an external electronic device. The notification relay application may, for example, transmit notification information corresponding to a designated event (e.g., receipt of an email) that occurred in another application (e.g., an email application (269)) of the electronic device (101) to the external electronic device. Additionally or alternatively, the notification relay application may receive notification information from the external electronic device and provide the information to the user of the electronic device (101).

[0068] A device management application may, for example, control the power (e.g., turning on or off) or the function (e.g., brightness, resolution, or focus) of an external electronic device or a component thereof (e.g., a display module or a camera module of the external electronic device) that communicates with the electronic device (101). The device management application may additionally or alternatively support the installation, deletion, or update of an application running on the external electronic device.

[0069] Throughout this specification, the terms "neural network," "neural network," and "network function" may be used interchangeably. A neural network may be comprised of a set of interconnected computational units, generally referred to as "nodes." These "nodes" may also be referred to as "neurons." A neural network comprises at least two nodes. The nodes (or neurons) comprising a neural network may be interconnected by one or more "links."

[0070] Within a neural network, two or more nodes connected via links can form a relationship between input and output nodes. The concepts of input and output nodes are relative, meaning that any node that is in an output relationship with one node can also be in an input relationship with another node, and vice versa. As described above, input-to-output node relationships can be created based on links. One input node can be connected to one or more output nodes via links, and vice versa.

[0071] In a relationship between input nodes and output nodes connected through a single link, the value of the output node can be determined based on the data input to the input node. Here, the node interconnecting the input nodes and output nodes can have a weight. The weight can be variable and can be varied by a user or an algorithm so that the neural network can perform a desired function. For example, when one or more input nodes are interconnected to one output node through each link, the output node can determine the output node value based on the values ​​input to the input nodes connected to the output node and the weight set for the link corresponding to each input node.

[0072] As described above, a neural network is a network in which two or more nodes are interconnected through one or more links, forming input and output node relationships within the network. The characteristics of a neural network can be determined based on the number of nodes and links within the network, the relationships between the nodes and links, and the weight values ​​assigned to each link. For example, if two neural networks have the same number of nodes and links but different weight values ​​between the links, the two neural networks can be perceived as different from each other.

[0073] FIG. 3 is a diagram of a system for providing a service for predicting the health index and disease of a companion animal using artificial intelligence, according to one embodiment. The embodiment of FIG. 3 can be combined with various embodiments of the present disclosure.

[0074] Referring to FIG. 3, a system for providing a service for predicting a pet's health index and disease through artificial intelligence (hereinafter, "prediction service") includes a service providing device (300) and a user terminal (400).

[0075] The service providing device (300) may be a device that provides a user terminal with information predicting the health index and disease of a companion animal through artificial intelligence. For example, the service providing device (300) may include the server (108) of FIG. 1.

[0076] The service providing device (300) can receive basic information about the companion animal from the user terminal (400). The service providing device (300) can receive inspection information about the companion animal from the user terminal (400). For example, the service providing device (300) can obtain inspection information about the companion animal matched based on the companion animal's basic information.

[0077] The service providing device (300) may provide a user interface (UI) for selecting or inputting basic information about a companion animal into the user terminal (400). The service providing device (300) may provide a user interface for inputting examination information about a companion animal into the user terminal (400). For example, the service providing device (300) may receive examination information about a companion animal, including at least one of a hair analysis result (hereinafter, “hair analysis result”) or a blood analysis result (hereinafter, “blood analysis result”) of the user's companion animal, through an input interface device included in the service providing device (300). For example, the service providing device (300) may receive examination information about a companion animal from a separate server.

[0078] The service providing device (300) can collect basic information about multiple companion animals, examination information about multiple companion animals, and veterinary diagnostic information about multiple companion animals. For example, the service providing device (300) can acquire basic information about multiple companion animals, examination information about multiple companion animals, and veterinary diagnostic information about multiple companion animals through internal data collection and refinement and external data web crawling collection and refinement. For example, the veterinary diagnostic information about a companion animal may include at least one of a veterinary opinion about the companion animal, a guardian's observations of the companion animal, multiple health indices for each item of the companion animal, a diagnosis of a disease, or the probability of disease onset. The veterinary opinion about the companion animal is the veterinarian's professional opinion about the companion animal and may be a comprehensive opinion about the companion animal's health status. The guardian's observations of the companion animal may be observations made by the guardian about the companion animal.

[0079] For example, the service providing device (300) may process error data and missing data regarding basic information of multiple companion animals, examination information of multiple companion animals, and veterinary diagnostic information for multiple companion animals before labeling is performed on data for training a machine learning model. For example, basic information of the companion animals, hair analysis results, and blood analysis test results may be labeled for multiple health indices by item. For example, basic information of the companion animals, hair analysis results, and blood analysis test results may be labeled for multiple diseases.

[0080] For example, multiple items may include circulatory (heart), cranial nerves, respiratory (lungs), ophthalmology, orthopedics (bone and joint), genitourinary, digestive, dental, endocrine, and skin. However, this is not limited to these, and various items may be set. For example, the health index may be a score of 1 or more and 5 or less. In this case, a lower health index score may indicate a worse health status for the corresponding item. For example, a score of 1 to 2 may correspond to the disease concern level, while a score of 3 to 5 may correspond to the health management level.

[0081] For example, multiple diseases may include diseases related to orthopedic disorders, diseases related to digestive disorders, diseases related to heart disease, diseases related to liver disease, diseases related to kidney disease, diseases related to skin diseases, and diseases related to urinary diseases. However, this is not limited to these diseases, and diseases related to various diseases may be set.

[0082] The service providing device (300) can determine multiple health indices of the pet by using a first machine learning model based on the pet's basic information and the pet's examination information.

[0083] The service providing device (300) can determine prediction information about the pet's disease by using a second machine learning model based on the pet's basic information and the pet's examination information.

[0084] The service providing device (300) can generate a result report including multiple health indices for each item of the companion animal and prediction information regarding the companion animal's disease. The service providing device (300) can transmit the result report to the user terminal (400).

[0085] The service provider (300) can use a neural network to determine a care method for the pet based on basic pet information, multiple health indices for the pet, and predicted disease information for the pet. For example, the result report may further include a care method for the pet.

[0086] The service providing device (300) can match a user terminal to one of N groups based on the basic information of the companion animal, the health index of the companion animal for each of multiple items, and the prediction information on the companion animal's disease.

[0087] The user terminal (400) is a terminal of a user who wishes to use a service for predicting the health index and disease of a companion animal, and transmits basic information of the user's companion animal and examination information of the companion animal to the service providing device (300), and can use various functions related to the service for predicting the health index and disease of the companion animal through the service providing device (300). For example, the companion animal's basic information may include the companion animal's age, sex, weight, neutering status, species, breed, and BCS (body condition score). The BCS may be a body condition index indicating the degree of obesity of the companion animal. For example, the BCS may include a value between 1 and 5. In this case, the BCS may include whether the ribs are visible to the naked eye, whether the waist and torso lines are clearly visible, the amount of subcutaneous fat, and

[0088] Muscle mass can be determined based on the amount of muscle mass. For example, if the ribs, spine, and lumbar vertebrae are visible, muscle mass is clearly absent, and there is no palpable subcutaneous fat, the pet may be thin, and the BCS may be assigned a value of 1. For example, if the ribs are easily visible, the lumbar curve is visible, and the belly curve rises from the front paws to the hind paws when observed from the side, the pet may be underweight, and the BCS may be assigned a value of 2. For example, if the lumbar curve is visible when observed from above, the ribs are palpable, and a small amount of subcutaneous fat is felt, the pet may be in ideal condition, and the BCS may be assigned a value of 3. For example, if the side view shows a slight curve from the front paws to the hind paws, but the pet has a wide waistline, fat on the abdomen and tail, and the ribs are palpable but there is a lot of subcutaneous fat, the pet may be overweight, and the BCS may be assigned a value of 4. For example, if your pet has no lumbar curvature, no bend from the front paws to the hind paws when viewed from the side, difficulty in palpating the ribs, and a thick layer of fat is present, the pet may be obese, and a BCS of 5 may be determined.

[0089] The pet's test information may include at least one of a hair analysis result or a blood analysis result. For example, the pet's test information may be obtained by the user terminal (400) requesting a test institution to analyze the user's pet's hair or blood. For example, the user terminal (400) may include the electronic device (101) of FIG. 1.

[0090] Hair analysis results may include values ​​for multiple hazardous heavy metals and values ​​for multiple nutritional minerals. For example, the multiple hazardous heavy metals may include at least one of mercury (Hg), arsenic (As), cadmium (Cd), lead (Pb), aluminum (Al), barium (Ba), uranium (U), bismuth (Bi), nickel (Ni), thyllium (Tl), cesium (Cs), or tin (Sb). For example, the multiple nutritional minerals may include at least one of calcium (Ca), magnesium (Mg), sodium (Na), potassium (K), copper (Cu), zinc (Zn), phosphorus (P), iron (Fe), manganese (Mn), chromium (Cr), selenium (Se), cobalt (Co), lithium (Li), vanadium (V), molybdenum (Mo), boron (B), or beryllium (Be).

[0091] Blood test results may include values ​​for a complete blood count and values ​​for serum biochemistry.

[0092] The values ​​for a complete blood count (CBC) may include values ​​related to red blood cells, values ​​related to white blood cells, and values ​​related to platelets. For example, values ​​related to red blood cells may include at least one of red blood cells, hematocrit, hemoglobin, mean corpuscular volume, hemoglobin per red blood cell, hemoglobin concentration per red blood cell, red blood cell size distribution width, reticulocytes, reticulocyte concentration, or reticulocyte hemoglobin. For example, values ​​related to white blood cells may include at least one of white blood cells, neutrophil concentration, lymphocyte concentration, monocyte concentration, eosinophil concentration, basophil concentration, neutrophils, lymphocytes, monocytes, eosinophils, or basophils. Values ​​related to platelets may include at least one of platelets, mean platelet volume, platelet particle size distribution width, or platelet volume percentage.

[0093] Serum biochemistry test values ​​may include values ​​related to liver function, kidney function, and diabetes. For example, liver function values ​​may include at least one of total protein, albumin, globulin, the albumin-to-globulin ratio, alkaline phosphatase (ALP), or alanine aminotransferase (ALT). Kidney function values ​​may include at least one of blood urea nitrogen (BUN), creatinine, or the ratio of BUN to creatinine. Diabetes values ​​may include glucose.

[0094] FIG. 4 is a block diagram exemplarily showing a functional module of a service providing device (300) according to FIG. 3. The embodiment of FIG. 4 can be combined with various embodiments of the present disclosure.

[0095] Referring to FIG. 4, the service providing device (300) may include a pet information receiving unit (310), a health index determining unit (320), a disease prediction unit (330), a care method recommendation unit (340), a group matching unit (350), and a report generating unit (360).

[0096] The companion animal information receiving unit (310) can receive basic information about the companion animal from the user terminal (400). The companion animal information receiving unit (310) can receive an age selected or entered within a preset range. For example, the companion animal's age can be set to 1 to 20 years old. However, the present invention is not limited thereto, and various ranges of ages can be set. The companion animal information receiving unit (310) can receive information about whether the companion animal is neutered. The user terminal (400) can select or input neutering or non-neutering through a user interface. The companion animal information receiving unit (310) can receive the companion animal's weight. For example, the companion animal's weight range can be set to 1 to 80 kg. However, the present invention is not limited thereto, and various ranges of weight can be set. In addition, units of weight can be kg, g, and lb. The companion animal information receiving unit (310) can receive a species selected from a preset selection. For example, the pet's breed may be selected as a dog or a cat. However, the present invention is not limited thereto, and various breeds may be set. The pet information receiving unit (310) may receive the pet's breed selected from preset options. For example, the pet's breed may be selected as any one of a plurality of preset breeds according to the species. The pet information receiving unit (310) may receive the value for the BCS of the pet selected from the preset options. For example, the value for the BCS may be selected from a value between 1 and 5.

[0097] The companion animal information receiving unit (310) can receive companion animal inspection information from the user terminal (400). The companion animal inspection information may include at least one of hair analysis results or blood analysis results. Alternatively, for example, the companion animal information receiving unit (310) can receive companion animal inspection information from an external server. The user terminal (400) can be configured to request a companion animal inspection from the external server and transmit the companion animal inspection information to the service providing device (300).

[0098] The health index determination unit (320) inputs the basic information of the companion animal and the examination information of the companion animal as input values ​​to a pre-learned first machine learning model, and can obtain multiple health indices of each item of the companion animal from the first machine learning model.

[0099] The first machine learning model may be machine-learned to output multiple health indices of a pet when inputting at least one of a pet hair analysis result or a pet blood analysis result and the pet's age, sex, neutering status, weight, species, breed, and BCS as input values. For example, the first machine learning model may output multiple health indices of a pet when inputting the pet's hair analysis result and the pet's age, sex, neutering status, weight, species, breed, and BCS as input values. For example, the first machine learning model may output multiple health indices of a pet when inputting the pet's blood analysis result and the pet's age, sex, neutering status, weight, species, breed, and BCS as input values. For example, the first machine learning model may output multiple health indices of a pet when inputting the pet's hair analysis result, the pet's blood analysis result, and the pet's age, sex, neutering status, weight, species, breed, and BCS as input values.

[0100] For training the first machine learning model, multiple first influential variables for each item can be used as training data for the first machine learning model. The multiple first influential variables for each item can be variables whose correlation score with the health index of the corresponding item is greater than or equal to a preset score among multiple first variables related to hair analysis results and multiple second variables related to blood analysis results. For example, random forests, Xgboost, multiple regression analysis, etc. can be used for training the first machine learning model.

[0101] FIG. 5 is a flowchart illustrating a process by which the health index determination unit (320) according to FIG. 4 determines multiple health indices for each item of a companion animal. The embodiment of FIG. 5 can be combined with various embodiments of the present disclosure.

[0102] Referring to FIG. 5, in step S510, the health index determination unit (320) may determine, among a plurality of first test variables for hair analysis results and a plurality of second test variables related to blood analysis results, a test variable whose correlation score for the health index is greater than or equal to a preset score as a first influence variable for each of a plurality of items.

[0103] For example, the correlation score may be based on at least one of the Pearson correlation coefficient or Spearman correlation coefficient. For example, the correlation score may be a value that measures the linear relationship between a test variable and a health index using the Pearson correlation coefficient. In this case, the correlation score may be determined based on the covariance of the test variable and the health index and the standard deviation of the variable and the health index.

[0104] For example, a correlation score can be a measure of the monotonic relationship between a test variable and a health index, using the Spearman correlation coefficient. In this case, the correlation score can be determined using the rank order of the test variable and the health index.

[0105] For example, a correlation score can be the absolute value of the Pearson correlation coefficient. For example, a correlation score can be the absolute value of the Spearman correlation coefficient. For example, a correlation score can be the average of the absolute values ​​of the Pearson correlation coefficient and the Spearman correlation coefficient.

[0106] Additionally, the preset score may be set to a value greater than 0.5 and less than 1. For example, the preset score may change depending on the performance of the first machine learning model. The preset score may be inversely proportional to the root mean square error (RMSE) value between the predicted value of the first machine learning model and the observed value. For example, the preset score may be set to a value closer to 0.5 as the root mean square error (RMSE) value between the predicted value of the first machine learning model and the observed value increases. For example, the preset score may be set to a value closer to 1 as the RMSE value between the predicted value of the first machine learning model decreases.

[0107] Multiple item-specific first-influence variables can be used as training data for the first machine learning model.

[0108] In step S520, the health index determination unit (320) can train a first machine learning model based on the basic variables for each of the plurality of companion animals, the first influence variables for each of the plurality of items, and the health indices for each of the plurality of items.

[0109] Here, basic variables may include age, gender, neutering status, species, breed, and BCS.

[0110] For example, a single baseline variable for a single companion animal, multiple item-specific first influence variables, and multiple item-specific health indices may be configured as a single first training data set for training a first machine learning model. For example, the first machine learning model may be trained using multiple first training data sets.

[0111] For example, ensemble techniques can be used as a training method for the first machine learning model. Ensemble techniques combine multiple individual models to generalize into an optimal model, offering the advantage of higher predictive performance than using a single model. Furthermore, ensemble techniques can address overfitting by training with weak classifiers, and can also address underfitting by combining predictions from multiple classifiers to improve performance.

[0112] For example, XGboost can be used to train the first machine learning model. At this time, the health index determination unit (320) can sequentially train multiple weak learners, and while predicting, it can assign weights to incorrectly predicted data to improve errors and gradually train the first machine learning model. In other words, based on the learning results of the previous model, a high weight can be assigned to an incorrect answer and a low weight can be assigned to a correct answer, and the assigned weights can influence the next model.

[0113] For example, the health index determination unit (320) may train a plurality of first machine learning models and select the first machine learning model having the smallest difference between the predicted value and the observed value. For example, the health index determination unit (320) may train a plurality of first machine learning models and select the first machine learning model having the smallest RMSE (root mean square error) value between the predicted value and the observed value. Here, the observed value

[0114] For example, the RMSE value can be determined by the following mathematical expression 1.

[0115]

[0116] In the above mathematical expression 1, k is the number of multiple items, y(i) is the observed value for the health index of the i-th item, can be a predicted value for the health index of the ith item.

[0117] In the above-described embodiment, the RMSE value can be replaced with the MAE (mean absolute error) value. Here, the MAE value is can be decided by

[0118] In step S530, the health index determination unit (320) can generate a basic vector through data preprocessing of basic information of the companion animal received from the user terminal.

[0119] The base vector may include values ​​for age, sex, neutering status, species, breed, and BCS. For example, the value for neutering status may be set to 1 if neutered and 0 if not neutered. For example, the value for species may be a value representing one of a plurality of species. For example, values ​​representing each of the plurality of species may be pre-stored in the service providing device (300). For example, the value for breed may be a value representing one of a plurality of breeds. For example, values ​​representing each of the plurality of breeds may be pre-stored in the service providing device (300).

[0120] In step S540, the health index determination unit (320) can generate a result vector through data preprocessing of the pet's examination information.

[0121] For example, the user terminal (400) can transmit the pet's examination information to the pet information receiving unit (310). Alternatively, for example, the pet information receiving unit (310) can receive the pet's examination information from an external server.

[0122] The result vector may include at least one of a plurality of values ​​associated with the hair analysis results or a plurality of values ​​associated with the blood analysis results.

[0123] For example, the plurality of values ​​associated with the hair analysis results may include numerical values ​​for a plurality of hazardous heavy metals and numerical values ​​for a plurality of nutritional minerals. The numerical values ​​for the plurality of hazardous heavy metals may include a value for at least one of mercury (Hg), arsenic (As), cadmium (Cd), lead (Pb), aluminum (Al), barium (Ba), uranium (U), bismuth (Bi), nickel (Ni), thyllium (Tl), cesium (Cs), or tin (Sb). The numerical values ​​for the plurality of nutritional minerals may include a value for at least one of calcium (Ca), magnesium (Mg), sodium (Na), potassium (K), copper (Cu), zinc (Zn), phosphorus (P), iron (Fe), manganese (Mn), chromium (Cr), selenium (Se), cobalt (Co), lithium (Li), vanadium (V), molybdenum (Mo), boron (B), or beryllium (Be).

[0124] For example, a plurality of values ​​associated with a blood analysis result may include a numerical value for a complete blood count and a numerical value for a serum biochemistry test. The numerical value for a complete blood count may include a numerical value associated with red blood cells, a numerical value associated with white blood cells, and a numerical value associated with platelets. The numerical value associated with red blood cells may include a value for at least one of red blood cells, hematocrit, hemoglobin, mean corpuscular volume, hemoglobin per red blood cell, hemoglobin concentration per red blood cell, red blood cell size distribution width, reticulocytes, reticulocyte concentration, or reticulocyte hemoglobin. For example, the numerical value associated with white blood cells may include a value for at least one of white blood cells, neutrophil concentration, lymphocyte concentration, monocyte concentration, eosinophil concentration, basophil concentration, neutrophils, lymphocytes, monocytes, eosinophils, or basophils. Numerical values ​​related to platelets may include values ​​for at least one of platelets, mean platelet volume, platelet particle distribution width, or platelet volume percentage. Numerical values ​​for serum biochemistry tests may include values ​​related to liver function, kidney function, and diabetes. For example, numerical values ​​related to liver function may include values ​​for at least one of total protein, albumin, globulin, albumin to globulin ratio, alkaline phosphatase (ALP), or alanine aminotransferase (ALT). Numerical values ​​related to kidney function may include values ​​for at least one of blood urea nitrogen, creatinine, or the ratio of blood urea nitrogen to creatinine. Numerical values ​​related to diabetes may include values ​​for glucose.

[0125] In step S550, the health index determination unit (320) can obtain multiple health indices for each item output from the first machine learning model by inputting the basic vector and the result vector into the first machine learning model.

[0126] For example, a health index can be determined by a score between 1 and 5. A lower score on the health index may indicate worse health for that specific item. For example, a score of 1 to 2 might correspond to a disease concern level, while a score of 3 to 5 might correspond to a health management level.

[0127] FIG. 6 is an example of an algorithm of the first machine learning model used in the health index determination unit (320) according to FIG. 4. FIG. 7 is an example of a learning result of the algorithm of the first machine learning model used in the health index determination unit (320) according to FIG. 4. The embodiments of FIGS. 6 and 7 can be combined with various embodiments of the present disclosure.

[0128] Referring to Figure 6, the first machine learning model can be configured with a boosting round (n_estimators) value of 50 or greater. Boosting rounds represent the number of trees to be generated. A higher value increases model complexity but also increases the risk of overfitting. Therefore, selecting a value that optimizes predictive performance may be important. For example, the boosting round value of the first machine learning model is preferably set to 100.

[0129] The first machine learning model can have a learning rate value set to 0.01 or greater and 0.3 or less. The learning rate indicates the degree to which the weights for each tree's predicted values ​​are updated. A small learning rate stabilizes the model, but may lead to overfitting. For example, the learning rate of the first machine learning model is preferably set to 0.08.

[0130] The first machine learning model can have a subsample value set to a value between 0.5 and 1.0. The subsample represents the proportion of the training data randomly selected when training each tree. While a small value increases the diversity of the trees, it can lead to overfitting. For example, the subsample value for the first machine learning model is preferably set to 0.75.

[0131] The first machine learning model can have a maximum depth (max_depth) value set to 3 or greater. The maximum depth represents the maximum depth of each tree. While larger values ​​allow for learning more complex patterns, they can also lead to overfitting. For example, the maximum depth value of the first machine learning model is preferably set to 7.

[0132] Additionally, for example, multiple first machine learning models may have different values ​​that can be designated as hyperparameters. For example, the health index determination unit (320) may sequentially input values ​​that can be designated as pre-determined hyperparameters into the first machine learning model, and then determine the hyperparameter with the highest performance (e.g., the lowest RSME value). In other words, the health index determination unit (320) may sequentially input values ​​that can be designated as hyperparameters and select the first machine learning model with the highest performance.

[0133] Alternatively, for example, the health index determination unit (320) may input a randomly selected value from among values ​​that can be designated as pre-determined hyperparameters into the first machine learning model, and then determine the hyperparameter with the highest performance (e.g., the lowest RSME value). That is, the health index determination unit (320) may select the first machine learning model with the highest performance while randomly inputting values ​​that can be designated as hyperparameters.

[0134] Alternatively, for example, the health index determination unit (320) may generate a table of values ​​that can be designated as predetermined hyperparameters. The table may be composed of multiple grids, and each grid may have a different hyperparameter. The health index determination unit (320) may randomly input hyperparameters corresponding to grids with predetermined intervals in the table, and determine candidate hyperparameters having RSME values ​​smaller than the predetermined RSME values. In addition, the health index determination unit (320) may determine a candidate region in the table in which candidate hyperparameters having RSME values ​​smaller than the predetermined RSME values ​​are densely concentrated, and the health index determination unit (320) may input all hyperparameters corresponding to grids included in the candidate region, and determine the hyperparameters having the highest performance (e.g., the lowest RSME value). This may reduce the number of unnecessary searches, and since numbers are entered randomly and then values ​​located between predetermined intervals (grids) can be probabilistically searched, the optimal hyperparameter values ​​may be found more quickly.

[0135] In the above-described embodiment, the RMSE value can be replaced with the mean absolute error (MAE) value.

[0136] Referring to Figure 7, it can be confirmed that the average MAE of the training dataset and the average MAE of the test dataset for multiple item-specific health indices of the first machine learning model are at acceptable levels.

[0137] Additionally, the health index determination unit (320) can determine multiple health indices for each item for training the first machine learning model based on veterinary opinions and guardian observations of the pets collected through internal data collection and refinement and external data web crawling collection and refinement for multiple pets.

[0138] Alternatively, for example, if the veterinary diagnostic information for a pet does not include multiple health indices, multiple health indices can be labeled by an external terminal based on the veterinary opinion, the pet owner's observations, and the diagnosis. The external terminal may be a terminal used by professionals such as veterinarians.

[0139] The disease prediction unit (330) inputs basic information and examination information about the companion animal as input values ​​to a pre-trained second machine learning model, and obtains prediction information about the companion animal's disease from the second machine learning model. For example, the prediction information about the companion animal's disease may include at least one disease or a healthy state.

[0140] The second machine learning model may be machine-learned to output at least one disease or health condition when inputting at least one of the pet's hair analysis result or the pet's blood analysis result and the pet's age, sex, neutering status, weight, species, breed, and BCS as input values. For example, the second machine learning model may output at least one disease or health condition when inputting the pet's hair analysis result and the pet's age, sex, neutering status, weight, species, breed, and BCS as input values. For example, the second machine learning model may output at least one disease or health condition when inputting the pet's blood analysis result and the pet's age, sex, neutering status, weight, species, breed, and BCS as input values. For example, the second machine learning model may output at least one disease or health condition when inputting the pet's hair analysis result, the pet's blood analysis result, and the pet's age, sex, neutering status, weight, species, breed, and BCS as input values.

[0141] For training the second machine learning model, second influential variables for each of the multiple diseases may be used as training data for the second machine learning model. The second influential variables for each of the multiple diseases may be variables among the multiple first variables related to the hair analysis results and the multiple second variables related to the blood analysis results, among which a correlation score for the corresponding disease is greater than or equal to a preset score. For example, random forests, Xgboost, multiple regression analysis, etc. may be used for training the second machine learning model.

[0142] FIG. 8 is a flowchart illustrating a process by which the disease prediction unit (330) according to FIG. 4 determines prediction information regarding a companion animal's disease. The embodiment of FIG. 8 can be combined with various embodiments of the present disclosure.

[0143] Referring to FIG. 8, in step S810, the disease prediction unit (330) may determine a test variable whose correlation score for the disease is higher than a preset score among a plurality of first test variables for the hair analysis results and a plurality of second test variables related to the blood analysis results as a second influence variable for each disease.

[0144] For example, the correlation score may be based on at least one of the Pearson correlation coefficient or the Spearman correlation coefficient. For example, the correlation score may be determined in the same manner as the correlation score in FIG. 5. The preset score may be set to a value greater than 0.5 and less than 1. For example, the preset score may change depending on the performance of the second machine learning model.

[0145] For each disease, a second influencing variable can be used as training data for a second machine learning model.

[0146] In step S820, the disease prediction unit (330) can train a second machine learning model based on the basic variables for each of the plurality of companion animals, the second influence variables for each of the plurality of diseases, and the plurality of diseases.

[0147] Here, basic variables may include age, gender, neutering status, species, breed, and BCS.

[0148] For example, an ensemble technique can be used as a learning method for the second machine learning model. For example, XGboost can be used to learn the second machine learning model. At this time, the disease prediction unit (330) can sequentially train multiple weak learners, and while predicting, it can gradually train the second machine learning model by assigning weights to incorrectly predicted data to improve errors. In other words, based on the learning results of the previous model, a high weight can be assigned to an incorrect answer and a low weight can be assigned to a correct answer, and the assigned weights can affect the next model.

[0149] Here, the disease may be a diagnosis for the companion animal included in the veterinary diagnostic information.

[0150] For example, a primary variable for a pet, a disease of the pet, and a secondary variable affecting that disease may be configured as a second training data set for training a second machine learning model. For example, the second machine learning model may be trained using multiple second training data sets.

[0151] For example, the disease prediction unit (330) may train a plurality of second machine learning models and select the second machine learning model based on the F1 score, recall, precision, and accuracy. For example, the disease prediction unit (330) may train a plurality of second machine learning models and select the second machine learning model in which the F1 score, recall, precision, and accuracy are greater than or equal to a preset threshold value, and the sum of the F1 score, recall, precision, and accuracy is the largest. The F1 score may be the harmonic mean of the recall and the precision. The recall may be an evaluation index that measures whether a classification has been made correctly. The precision may be an evaluation index that measures whether a prediction has been made correctly. The accuracy may indicate the proportion of samples that have been classified correctly among all predictions.

[0152] For example, the F1 score can be determined by the following mathematical expression 2.

[0153]

[0154] In the above mathematical expression 2, Precision may be precision, and Recall may be recall.

[0155] For example, the precision can be determined by the following mathematical expression 3.

[0156]

[0157] In the above mathematical expression 3, TP may be the number of samples correctly predicted as positive (True), and FP may be the number of samples that are negative (False) but predicted as positive (True).

[0158] For example, the recall rate can be determined by the following mathematical equation 4.

[0159]

[0160] In the above mathematical expression 4, TP may be the number of samples correctly predicted as positive, and FN may be the number of samples that are positive but predicted as negative.

[0161] For example, the precision can be determined by the following mathematical expression 5.

[0162]

[0163] In the above mathematical expression 5, TP may be the number of samples correctly predicted as positive, TN may be the number of samples correctly predicted as negative, FP may be the number of samples that are negative but predicted as positive, and FN may be the number of samples that are positive but predicted as negative.

[0164] In step S830, the disease prediction unit (330) can generate a basic vector through data preprocessing of basic information of the companion animal received from the user terminal.

[0165] In step S840, the disease prediction unit (330) can generate a result vector through data preprocessing of the pet's examination information.

[0166] In step S850, the disease prediction unit (330) can obtain at least one disease or healthy state among multiple diseases output from the second machine learning model by inputting the base vector and the result vector into the second machine learning model.

[0167] For example, the disease prediction unit (330) may include at least one disease or health condition in the prediction information about the pet's disease.

[0168] FIG. 9 is an example of an algorithm of the second machine learning model used in the disease prediction unit (330) according to FIG. 4. FIG. 10 is an example of a learning result of the algorithm of the second machine learning model used in the disease prediction unit (330) according to FIG. 4. FIG. 14 is another example of a learning result of the algorithm of the second machine learning model used in the disease prediction unit (330) according to FIG. 4. The embodiments of FIGS. 9, 10, and 14 can be combined with various embodiments of the present disclosure.

[0169] Referring to Figure 9, the second machine learning model can set missing values ​​to 1. Missing values ​​can indicate empty or nonexistent values ​​in the data. For example, by setting missing values ​​to 1, the second machine learning model can learn to reflect values ​​indicating a healthy state.

[0170] The second machine learning model can have a learning rate value greater than 0.01 and less than 0.3. The learning rate indicates the degree to which the weights for each tree's predicted values ​​are updated. A small learning rate stabilizes the model, but may lead to overfitting. For example, the learning rate of the second machine learning model is preferably set to 0.1.

[0171] The second machine learning model can have a maximum depth (max_depth) value of 3 or greater. The maximum depth represents the maximum depth of each tree. While larger values ​​allow for learning more complex patterns, they can also lead to overfitting. For example, the maximum depth value of the second machine learning model is preferably set to 7.

[0172] The second machine learning model can have an early stopping round value of 5 or more and 20 or less. The early stopping round value can be the number of times training is stopped early if performance does not improve after a certain number of boosting rounds. For example, the early stopping round value for the second machine learning model is preferably set to 10.

[0173] Additionally, for example, multiple second machine learning models may have different values ​​that can be designated as hyperparameters. For example, the disease prediction unit (330) may sequentially input values ​​that can be designated as predetermined hyperparameters into the second machine learning model and then determine the hyperparameter with the highest performance. For example, the highest performance in the second machine learning model may be one in which the F1 score, recall, precision, and accuracy are above a preset threshold value, and the sum of the F1 score, recall, precision, and accuracy is the largest. In other words, the disease prediction unit (330) may sequentially input values ​​that can be designated as hyperparameters and select the second machine learning model with the highest performance.

[0174] Alternatively, for example, the disease prediction unit (330) may input randomly selected values ​​from among values ​​that can be designated as pre-determined hyperparameters into the second machine learning model, and then determine the hyperparameter with the highest performance. In other words, the disease prediction unit (330) may select the first machine learning model with the highest performance while randomly inputting values ​​that can be designated as hyperparameters.

[0175] Alternatively, for example, the disease prediction unit (330) may generate a table of values ​​that can be specified as predetermined hyperparameters. The table may be composed of multiple grids, and each grid may have different hyperparameters. The disease prediction unit (330) may randomly input hyperparameters corresponding to grids with predetermined intervals within the table, and determine candidate hyperparameters having values ​​with greater F1 score, recall, precision, and accuracy than a predetermined reference value. In addition, the disease prediction unit (330) may determine a candidate region in which candidate hyperparameters are densely concentrated within the table, and the disease prediction unit (330) may input all hyperparameters corresponding to grids included in the candidate region, and determine the hyperparameter with the highest performance (e.g., the hyperparameter with the largest sum of F1 score, recall, precision, and accuracy). This may reduce the number of unnecessary searches, and since numbers are entered randomly and then values ​​located between predetermined intervals (grids) can be probabilistically searched, the optimal hyperparameter values ​​may be found more quickly.

[0176] Referring to Fig. 10, it can be confirmed that the F1 score, recall, precision, and accuracy of the second machine learning model's disease prediction results are at a satisfactory level. Specifically, Fig. 10 may be the test results for the second machine learning model that learned from a dataset of 1,800 cases consisting of veterinary diagnostic tests, hair tests, and blood tests, classifying disease codes into 0 (health), 1 (orthopedic disease), 2 (digestive disease), 3 (heart disease), 4 (tumor), 5 (kidney disease), 6 (skin disease), and 7 (other). At this time, the test confirmed the F1 score based on the correct answers and predicted values ​​for 400 test data sets.

[0177] Referring to Figure 14, it can be confirmed that the average AUC (Area Under the ROC Curve) value for the results of the second machine learning model predicting the disease is at a satisfactory level.

[0178] For example, the first and second machine learning models can use Classification And Regression Trees (CART) to generate decision trees. Since all leaf nodes of CART are related to the model's score, models with identical classification results can be compared for score comparison.

[0179] For example, the first machine learning model and the second machine learning model can include regularization to prevent overfitting. For example, the first machine learning model and the second machine learning model can penalize model complexity by adding a regularization term to the objective function. The regularization term can be L1 regularization and L -2 Regularization can be included. L1 regularization (Lasso) can add a penalty proportional to the sum of the absolute values ​​of the weights. This can achieve the effect of feature selection by making some weights exactly 0. L2 regularization (Ridge) can add a penalty proportional to the sum of the squares of the weights. This can keep weights small, while still bringing them closer to 0, without making them completely 0.

[0180] For example, the objective function can be determined by the following mathematical expression 6.

[0181]

[0182] In the above mathematical expression 6, the above Loss can represent the loss of the corresponding model, and the above can represent the strength of normalization, and the above is an L1 regularization term, which can represent the sum of the absolute values ​​of the weights, and the above is the L2 regularization term, which can be the sum of the squares of the weights, and can be a parameter that adjusts the weight between the L1 regularization term and the L2 regularization term.

[0183] For example, the strength of regularization can constrain the weights of a given module. For example, the regularization term for L1 regularization can have the effect of feature selection by setting some weights to zero. For example, the regularization term for L2 regularization can reduce weights and prevent overfitting. For example, can have a value greater than or equal to 0 and less than or equal to 1.

[0184] For example, the first machine learning model and the second machine learning model can reduce the number of splits by deleting splits that have no benefit (splittings with an information gain less than 0) through tree pruning.

[0185] For example, the first machine learning model and the second machine learning model have cross-validation built in, so they can perform cross-validation internally for each iteration to have an optimized number of iterations, and if the evaluation value of the evaluation dataset is optimized through cross-validation rather than a specified number of iterations, the iteration can be stopped in the middle.

[0186] The care method recommendation unit (340) can determine a recommended care method for a pet using a neural network based on the pet's basic information, multiple health indices for each item, and prediction information on the pet's disease.

[0187] FIG. 11 is a flowchart illustrating the process by which the care method recommendation unit (350) according to FIG. 4 determines a care method for a companion animal using a neural network. The embodiment of FIG. 11 can be combined with various embodiments of the present disclosure.

[0188] Referring to FIG. 11, in step S1110, the care method recommendation unit (350) can obtain multiple care methods from multiple user terminals that have received two or more result reports.

[0189] Here, the care method can refer to information on how to care for your pet. For example, care methods can include various types of care methods, such as the number of walks, distance, duration, prohibited foods, recommended foods and feeding amounts, food bowl material, types and materials of pet toys, and the material of the pet tag.

[0190] For example, a user terminal that receives two or more result reports may transmit a care method for the pet to the service providing device. For example, the care method may include at least one of the number of walks, the distance walked, the time of the walk, information on foods prohibited from consumption, information on foods recommended for consumption, the material of the food bowl, the type and material of the pet's toys, and the material of the pet tag. The number of walks may refer to the number of times the pet walks in a day. The distance walked may refer to the distance the pet walks in a day. The time walked may refer to the time the pet walks in a day. The prohibited foods may refer to the types of foods that are not fed to the pet. The recommended foods and daily feeding amount may refer to the types of foods fed to the pet, the number of times the food is fed in a day, and the amount fed at a time.

[0191] For example, if a user terminal receives a first result report and then, after a specific period of time, receives a second result report, the user terminal may transmit a care method for the pet to the service providing device. In other words, the first result report may be a result report received from the service providing device before the user terminal receives the second result report from the service providing device.

[0192] In step S1120, the care method recommendation unit (350) can determine a candidate care method for training a neural network among multiple care methods.

[0193] For example, the care method recommendation unit (350) may determine a care method among multiple care methods in which the multiple item-specific health indices included in the second result report are higher than the multiple item-specific health indices included in the first result report as a candidate care method for training a neural network.

[0194] In step S1130, the care method recommendation unit (350) can train a neural network based on basic information of multiple companion animals, multiple sets of health indices, prediction information on diseases of multiple companion animals, and multiple candidate care methods.

[0195] Here, the neural network may be a CNN-based neural network. For example, the neural network may include an input layer, one or more hidden layers, and an output layer.

[0196] Basic information of multiple companion animals, multiple sets of health indices, prediction information on diseases of multiple companion animals, and multiple candidate care methods may be pre-stored in the service providing device (300) to train a neural network. For example, a basic vector may be generated through data preprocessing for each of the basic information of multiple companion animals. For example, a health index vector may be generated through data preprocessing for each of the multiple sets of health indices. A health index set may refer to multiple health indices for each item of a single companion animal. The health index vector may include values ​​for each of the multiple health indices for each item. For example, a disease vector may be generated by performing data preprocessing on each of the prediction information on diseases of multiple companion animals. The disease vector may include either a value indicating at least one disease or a value indicating a healthy state.

[0197] For example, a user terminal that transmits a candidate care method can be classified into one of multiple care types based on the basic vector, health index vector, and disease vector.

[0198] For example, multiple care types can be determined through a clustering technique using multiple type vectors including the base vector, health index vector, and disease vector of the user terminal that transmitted the candidate care method. Here, clustering can refer to unsupervised learning that groups data with similar attributes into a certain number of clusters. Specifically, the multiple type vectors can be reduced to vectors of three dimensions or less through various dimensionality reduction techniques. For example, the dimensionality of the multiple type vectors can be reduced to a dimension of three dimensions or less through various dimensionality reduction techniques. For example, the type vectors can be reduced to a dimension of three dimensions or less through the principal component analysis (PCA) technique. For example, the care method recommendation unit (350) can determine the axis of data with the highest variance when projecting the multiple type vectors onto the principal component axis, and reduce the dimensionality to the determined axis. For example, the care method recommendation unit (350) can generate the first axis based on the largest variance among the plurality of type vectors, and can generate the second axis as a vector that is orthogonal to the first vector axis. Thereafter, the server can generate the third axis as a vector that is orthogonal to the second axis. When the care method recommendation unit (350) projects the original data onto the three generated vector axes, the care method recommendation unit (350) can reduce the dimensionality of the original data to the number of vector axes. Hereinafter, a plurality of type vectors are generated, and vectors that are obtained by reducing the dimensionality of the generated type vectors through various dimensionality reduction techniques may be referred to as first dimension reduction vectors. For example, a plurality of care types may be determined as n care types through the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) technique based on a plurality of first dimension reduction vectors.For example, DBSCAN assumes that if a specific element (point) belongs to a cluster, it must be located close to many other elements within the cluster, and for this calculation, a radius and minimum points can be used. For example, a diameter can be a radius based on a specific data element, which can be referred to as a dense area. For example, a minimum point can indicate how many elements are required around a core element to designate a core point. In addition, each element of a data set can be classified into a core, a border, and an outlier point. For example, the care method recommendation unit (350) can check the size of the diameter for each element and search for the number of surrounding elements. Thereafter, the care method recommendation unit (350) can determine an element as a core element if there are k or more elements within the diameter range. In addition, the care method recommendation unit (350) can determine an element included within the diameter range from the core element as a border element. Furthermore, the care method recommendation unit (350) can determine elements not included within the diameter range of the core elements as outlier elements, and these outlier elements can be excluded from the corresponding cluster. Furthermore, the care method recommendation unit (350) can classify the elements into the same cluster if the distance between the core elements is smaller than the diameter.

[0199] For example, training data consisting of multiple base vectors, multiple health index vectors, multiple disease vectors, and multiple correct answer type vectors is input to an input layer, passes through one or more hidden layers and an output layer, and is output as an output vector. The output vector is input to a loss function layer connected to the output layer. The loss function layer outputs a loss value using a loss function that compares the output vector with the correct answer vector for each training data, and the parameters of the neural network can be trained in a direction in which the loss value decreases.

[0200] One or more hidden layers may include one or more convolutional layers and one or more pooling layers. For example, multiple base vectors, multiple health index vectors, and multiple disease vectors may be filtered in the convolutional layers, and a feature map may be formed through the convolutional layers.

[0201] For example, by selecting a fixed vector related to a feature for dimensionality reduction based on a feature map formed in a pooling layer and performing sub-sampling on the formed feature map, a feature related to a type vector can be extracted from the vectorized time series data. For example, the pooling layer can be a max pooling layer that extracts the largest value. For example, the pooling layer can be an average pooling layer that extracts an average value. For example, at this time, the parameters of the neural network can include parameters related to the convolutional layer and the pooling layer (size of the feature map, size of the filter, depth, stride, zero padding).

[0202] The correct answer type vector may include the base vector, health index vector, and disease vector of the user terminal that transmitted the candidate care method. In this case, the correct answer type vector for each of the multiple user terminals that transmitted the candidate care method may be pre-stored in the service providing device for neural network learning.

[0203] This allows the neural network to learn to determine the most similar type vector based on the input base vector, health score vector, and disease vector.

[0204] In step S1140, the care method recommendation unit (350) can determine a care type that includes the output type vector among multiple care types by inputting the basic vector, health index vector, and disease vector into the neural network.

[0205] Here, the basic vector, health index vector, and disease vector can be generated based on basic information, multiple item-specific health indices, and prediction information on the companion animal's disease obtained from the user terminal that will provide the result report.

[0206] In step S1150, the care method recommendation unit (350) can determine the care method with the greatest improvement in health index among multiple candidate care methods included in the care type as the recommended care method.

[0207] The degree of improvement in the health index may be the sum of the increase in the multiple item-specific health indices included in the second result report compared to the multiple item-specific health indices included in the first result report.

[0208] For example, the care method recommendation unit (350) can transmit a recommended care method to the user terminal.

[0209] For example, the results report may further include recommended care methods for your pet.

[0210] The group matching unit (350) can match a user terminal to one of N groups based on the pet's basic information, multiple health indices for each item, and predicted information on the pet's disease.

[0211] For example, the group matching unit (350) can match a user terminal to one of N groups within a care type including the user terminal based on the location information of the user terminal.

[0212] For example, the group matching unit (350) can create N groups for one care type based on location information for multiple user terminals included in one care type.

[0213] For example, each group may be composed of user terminals located within a preset range. For example, the preset range may be determined such that each of the N groups satisfies a number of user terminals greater than or equal to a preset minimum value. In this case, for example, if the number of user terminals included in one group is less than the preset minimum value, the group may be merged into another group set at the location coordinates closest to the location coordinates set for the group. The location coordinates set for the group may be barycentric coordinates calculated based on the location coordinates of the user terminals included in the group. The location coordinates may be determined based on either GPS information of the user terminal or address information directly input by the user terminal. The location coordinates may be latitude and longitude included in the GPS information or latitude and longitude matching the address information directly input by the user terminal.

[0214] For example, the group matching unit (350) may transmit information about a group including a user terminal to the user terminal. The information about the group including the user terminal may include a value for the care type and the IDs of other user terminals included in the group. Here, the ID may be the ID with which the user terminal subscribed to the prediction service.

[0215] The report generation unit (360) can automatically generate a result report including multiple health indices for each item of the companion animal and prediction information on the companion animal's disease.

[0216] The report generation unit (360) can determine the probability of occurrence of at least one disease in a companion animal by using a neural network based on a health index vector and a disease vector.

[0217] Here, the neural network may be a CNN-based neural network. For example, the neural network may include an input layer, one or more hidden layers, and an output layer.

[0218] Multiple health index vectors, multiple disease vectors, and onset probabilities for multiple diseases may be pre-stored in the service providing device (300) to train a neural network. For example, a health index vector may be generated through data preprocessing for each set of multiple health indices. For example, a disease vector may be generated by performing data preprocessing on each prediction information for diseases of multiple companion animals. Each health index vector and each disease vector may be matched with the onset probabilities for multiple diseases and used as training data. In this case, the onset probabilities for multiple diseases may be data labeled by a veterinarian.

[0219] For example, training data consisting of multiple health index vectors, multiple disease vectors, and correct occurrence probabilities for multiple diseases is input to an input layer, passes through one or more hidden layers and an output layer, and outputs an output vector. The output vector is input to a loss function layer connected to the output layer. The loss function layer outputs a loss value using a loss function that compares the output vector with the correct answer vector for each training data. The parameters of the neural network can be trained in a direction in which the loss value decreases.

[0220] One or more hidden layers may include one or more convolutional layers and one or more pooling layers. For example, multiple health index vectors and multiple disease vectors may be filtered in the convolutional layer, and a feature map may be formed through the convolutional layer.

[0221] For example, by selecting fixed vectors related to features for dimensionality reduction based on the feature map formed in the pooling layer and performing sub-sampling on the formed feature map, features related to the onset probability of a disease can be extracted from the vectorized time series data. For example, the pooling layer can be a max pooling layer that extracts the largest value. For example, the pooling layer can be an average pooling layer that extracts an average value. For example, at this time, the parameters of the neural network can include parameters related to the convolutional layer and the pooling layer (size of the feature map, size of the filter, depth, stride, zero padding).

[0222] For example, a single health index vector and a disease vector can comprise a single training data set and the correct incidence probabilities for multiple diseases. Multiple training data sets can be pre-saved. The correct incidence probabilities can be labeled by a veterinarian.

[0223] Through this, the neural network can be trained to determine the probability of developing multiple diseases based on input of multiple health index vectors and multiple disease vectors.

[0224] For example, the report generation unit (360) can input a health index vector and a disease vector into a neural network to determine the probability of onset of multiple diseases.

[0225] For example, the report generation unit (360) can generate a result report including a plurality of health indices for each item of the companion animal, an average value of the plurality of health indices for each item of the companion animal, and an incidence probability for at least one disease of the companion animal.

[0226] For example, when a recommended care method is determined, the report generation unit (360) can generate a result report including the pet's multiple item-specific health indices, the average value of the pet's multiple item-specific health indices, the probability of onset of at least one disease of the pet, and the recommended care method.

[0227] Fig. 12 illustrates an example of a result report generated by the report generation unit according to Fig. 4. The embodiment of Fig. 12 can be combined with various embodiments of the present disclosure.

[0228] Referring to Figure 12, the report generation unit (360) can automatically generate a result report by applying a preset template to the pet's multiple health indices, the average of the multiple health indices, and the probability of the pet developing at least one disease. For example, the preset template may include a predefined design, layout, color, and font.

[0229] For example, the result report may be displayed differently depending on the website used by the user terminal, the mobile site used by the user terminal, the animal hospital-specific website, and the dedicated administrator page.

[0230] Fig. 13 is a block diagram illustrating the configuration of a service providing device according to one embodiment. The embodiment of Fig. 13 can be combined with various embodiments of the present disclosure.

[0231] As illustrated in FIG. 13, the service providing device (300) may include a processor (1310), a communication unit (1320), and a memory (1330). However, not all of the components illustrated in FIG. 13 are essential components of the service providing device (300). The service providing device (300) may be implemented with more components than the components illustrated in FIG. 13, or may be implemented with fewer components than the components illustrated in FIG. 13. For example, the service providing device (300) according to some embodiments may further include a user input interface (not illustrated), an output unit (not illustrated), etc., in addition to the processor (1310), the communication unit (1320), and the memory (1330).

[0232] The processor (1310) typically controls the overall operation of the service providing device (300). The processor (1310) may include one or more processors and control other components included in the service providing device (300). For example, the processor (1310) may control the communication unit (1320) and the memory (1330) by executing programs stored in the memory (1330). In addition, the processor (1310) may perform the functions of the service providing device (300) described in FIGS. 3 to 9 by executing programs stored in the memory (1330).

[0233] The communication unit (1320) may include one or more components that enable the service providing device (300) to communicate with other devices (not shown) and servers (not shown). The other devices (not shown) may be computing devices such as the service providing device (300) or sensing devices, but are not limited thereto. The communication unit (1320) may receive user input from other electronic devices or receive data stored in an external device from an external device via a network.

[0234] For example, the communication unit (1320) can transmit and receive messages for establishing a connection with at least one device. The communication unit (1320) can transmit information generated by the processor (1310) to at least one device connected to the management server. The communication unit (1320) can receive information from at least one device connected to the management server. The communication unit (1320) can transmit information related to the received information in response to the information received from at least one device.

[0235] The memory (1330) can store programs for processing and controlling the processor (1310). For example, the memory (1330) can store information input to the server or information received from another device via a network. In addition, the memory (1330) can store data generated by the processor (1310). The memory (1330) can also store information input to or output from the service providing device (300).

[0236] The memory (1330) may include at least one type of storage medium among a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (e.g., SD or XD memory, etc.), a RAM (Random Access Memory), a SRAM (Static Random Access Memory), a ROM (Read-Only Memory), an EEPROM (Electrically Erasable Programmable Read-Only Memory), a PROM (Programmable Read-Only Memory), a magnetic memory, a magnetic disk, and an optical disk.

[0237] The embodiments described above may be implemented using hardware components, software components, and / or a combination of hardware components and software components. For example, the devices, methods, and components described in the embodiments may be implemented using one or more general-purpose computers or special-purpose computers, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing instructions and responding to them. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing device is sometimes described as being used alone; however, one of ordinary skill in the art will recognize that the processing device may include multiple processing elements and / or multiple types of processing elements. For example, a processing unit may include multiple processors, or a processor and a controller. Other processing configurations, such as parallel processors, are also possible.

[0238] Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing device to perform a desired operation or may, independently or collectively, command the processing device. The software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave, for interpretation by the processing device or for providing instructions or data to the processing device. The software may also be distributed over networked computer systems and stored or executed in a distributed manner. The software and data may be stored on one or more computer-readable recording media.

[0239] The method according to the embodiment may be implemented in the form of program commands that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program commands, data files, data structures, etc., alone or in combination. The program commands recorded on the medium may be those specially designed and configured for the embodiment or may be those known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program commands, such as ROMs, RAMs, and flash memories. Examples of the program commands include not only machine language codes generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiment, and vice versa.

[0240] Although the embodiments described above have been described with limited drawings, those skilled in the art will appreciate that various technical modifications and variations can be applied based on the above. For example, appropriate results can still be achieved even if the described techniques are performed in a different order than described, and / or components of the described systems, structures, devices, circuits, etc. are combined or combined in a different manner than described, or are replaced or substituted with other components or equivalents.

[0241] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims described below.

Claims

1. A device that provides a service for predicting the health index and disease of companion animals through artificial intelligence. A pet information receiving unit that receives basic information about the pet and examination information about the pet; A health index determination unit that determines multiple health indices of the companion animal by using a first machine learning model based on the basic information of the companion animal and the examination information of the companion animal; A disease prediction unit that determines prediction information about the disease of the companion animal using a second machine learning model based on the basic information of the companion animal and the examination information of the companion animal; and Including a report generation unit that automatically generates a result report including multiple health indices for each item of the companion animal and prediction information on diseases of the companion animal, The above result report is transmitted to the user terminal, Among the plurality of first test variables for the hair analysis results and the plurality of second test variables related to the blood analysis results, a test variable having a correlation score for the health index greater than or equal to a preset score is determined as a first influential variable for each of the plurality of items, and the first influential variable is used as learning data for the first machine learning model. Among the plurality of first test variables and the plurality of second test variables, a test variable whose correlation score for a disease is greater than or equal to a preset score is determined as a second influential variable for each disease, and the second influential variable is used as learning data for the second machine learning model. method.

2. In paragraph 1, Through data preprocessing of the above basic information, a basic vector containing values ​​for each basic variable is created. The above basic variables include age, sex, neutering, species, breed, and BCS (body condition score). Through data preprocessing of the above test information, a result vector including at least one of a plurality of values ​​related to hair analysis results or a plurality of values ​​related to blood analysis results is generated, Based on the above basic vector and the above result vector being input to the first machine learning model, the above multiple item-specific health indices are output, The above first machine learning model is learned based on the basic variables for each of the plurality of companion animals, the first influence variables for each of the plurality of items, and the health indices for each of the plurality of items. device.

3. In paragraph 2, Based on the above base vector and the above result vector being input to the second machine learning model, at least one of a disease or a healthy state is output, The prediction information about the disease of the companion animal includes at least one of the disease or the healthy state, The above second machine learning model is learned based on the basic variables for each of the multiple companion animals, the second influence variables for each of the multiple diseases, and the multiple diseases. device.

4. In paragraph 3, Further including a care method recommendation section that determines a recommended care method for the companion animal using a neural network based on the basic information of the companion animal, the health index of each of the multiple items, and the prediction information on the disease of the companion animal. The above results report further includes recommended care methods for the above companion animal, The above neural network is trained based on basic information of multiple companion animals, multiple sets of health indices, prediction information on diseases of multiple companion animals, and multiple care methods. The above multiple care methods are obtained from multiple user terminals that have received two or more result reports, Among the above multiple care methods, a care method in which multiple item-specific health indices included in the second result report are higher than multiple item-specific health indices included in the first result report received before the second result report is determined as a candidate care method for training the neural network. device.

5. In paragraph 4, Further comprising a group matching unit that matches the user terminal to one of N groups based on the basic information of the companion animal, the health index of each of the plurality of items, and the prediction information on the disease of the companion animal. Multiple care types are determined through a clustering technique using multiple type vectors including the basic vector, health index vector, and disease vector of the user terminal that transmitted the candidate care method, The above health index vector contains values ​​for multiple health indices for each item, The disease vector comprises either a value representing at least one disease or a value representing a healthy state, The above N groups are generated for one care type based on the location information of the plurality of user terminals included in one care type. device.

Images