PROGRAM, MODEL GENERATION METHOD, INFORMATION PROCESSING METHOD, AND INFORMATION PROCESSING APPARATUS

Abstract

To provide a program or the like capable of easily determining a type of a headache.SOLUTION: A program according to an embodiment of the present disclosure acquires answer information of a user regarding headaches, and in a case where the answer information is input, the program causes a computer to execute a process of outputting classification information by inputting the acquired answer information to a learning model that has been trained to output the classification information indicating any of migraine, medication overuse headache, tension-type headache, trigeminal autonomic headache, and other primary headaches or secondary headaches.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present technology relates to a program, a model generation method, an information processing method, and an information processing device. [Background technology]

[0002] Diagnostic kits for diagnosing migraine have been proposed in the past. For example, the diagnostic kit described in Patent Document 1 diagnoses whether a subject suffers from migraine by examining the expression level of the α-fodrin gene or HPCAL1 gene in lymphocytes isolated from the subject. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 4925063 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the diagnostic kit described in Patent Document 1 requires the preparation of a cell extract from the subject's cells, and has the problem that the type of headache cannot be easily determined.

[0005] The present disclosure has been made in consideration of the above circumstances, and aims to provide a program, etc. that can easily determine the type of headache. [Means for solving the problem]

[0006] A program according to one embodiment of the present disclosure acquires user response information regarding headaches, and inputs the acquired response information into a learning model trained to output classification information indicating either a primary headache, including migraine, or a secondary headache, when response information is input, thereby causing a computer to execute a process of outputting the classification information.

[0007] A model generation method according to one embodiment of the present disclosure acquires training data including a user's response information regarding headaches and classification information indicating either a primary headache or a secondary headache including migraine determined based on the response information, and generates a learning model that outputs the classification information when the response information is input based on the acquired training data.

[0008] An information processing method according to one embodiment of the present disclosure acquires user response information regarding headaches, and outputs the classification information by inputting the acquired response information into a learning model trained to output classification information indicating either a primary headache, including migraine, or a secondary headache when response information is input.

[0009] An information processing device according to one embodiment of the present disclosure includes a control unit that acquires user response information regarding headaches, and outputs the classification information by inputting the acquired response information into a learning model that has been trained to output classification information indicating either a primary headache or a secondary headache, including migraine, when response information is input. Effect of the Invention

[0010] A program according to an embodiment of the present disclosure makes it possible to easily determine the type of headache. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 is a schematic diagram showing an overview of an information processing system. [Diagram 2] 1 is a block diagram showing an example of the configuration of an information processing device according to a first embodiment. [Diagram 3] FIG. 2 is a block diagram showing a configuration example of a terminal device. [Figure 4] FIG. 2 is an explanatory diagram showing an example of a learning model according to the first embodiment. [Diagram 5] FIG. 11 is an explanatory diagram illustrating an example of a diagnosis information table. [Figure 6] FIG. 11 is an explanatory diagram illustrating an example of a classification information history table. [Figure 7]FIG. 13 is an explanatory diagram showing an example of a medical questionnaire screen on the terminal device. [Figure 8] FIG. 13 is an explanatory diagram showing an example of a medical questionnaire screen on the terminal device. [Figure 9] FIG. 13 is an explanatory diagram showing an example of a diagnosis result display screen on the terminal device. [Figure 10] 5 is a flowchart showing an example of processing of the information processing device and the terminal device according to the first embodiment. [Figure 11] FIG. 11 is a block diagram showing an example of the configuration of an information processing device according to a second embodiment. [Figure 12] FIG. 13 is an explanatory diagram showing an example of a headache history table. [Figure 13] FIG. 11 is an explanatory diagram showing an example of a learning model according to the second embodiment. [Figure 14] 11 is a flowchart showing an example of processing of an information processing device and a terminal device according to the second embodiment. [Figure 15] FIG. 11 is a block diagram showing an example of the configuration of a terminal device according to a third embodiment. [Figure 16] FIG. 13 is an explanatory diagram showing an example of a headache record table. [Figure 17] FIG. 13 is an explanatory diagram showing an example of a medication record table. [Figure 18] FIG. 13 is an explanatory diagram showing an example of a headache record selection screen. [Figure 19] FIG. 13 is an explanatory diagram showing an example of a medication record selection screen. [Figure 20] 11 is a flowchart showing an example of processing of an information processing device and a terminal device according to a third embodiment. [Figure 21] FIG. 11 is a block diagram showing an example of the configuration of an information processing device according to a fourth embodiment. [Figure 22] FIG. 13 is an explanatory diagram illustrating an example of a contribution degree table. [Diagram 23] FIG. 13 is an explanatory diagram showing an example of a contribution graph display screen. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] (Embodiment 1) FIG. 1 is a schematic diagram showing an overview of an information processing system. The information processing system S includes an information processing device 1 and a terminal device 2. The information processing device 1 can communicate with the terminal device 2 via a network N using wide-area wireless communication. The terminal device 2 is a terminal owned by a user (headache patient). The terminal device 2 may be a terminal owned by a hospital or a doctor who examines headache patients. The terminal device 2 transmits answer information regarding the headache suffered by the user to the information processing device 1. Based on the answer information received from the terminal device 2, the information processing device 1 outputs classification information (diagnosis result) indicating the type of headache, and transmits the output classification information to the terminal device 2.

[0013] FIG. 2 is a block diagram showing a configuration example of the information processing device 1 according to the first embodiment. The information processing device 1 is, for example, a server computer, and includes a control unit 11, a storage unit 12, and a communication unit 13. The control unit 11 is configured with a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a GPU (Graphical Processing Unit), a quantum processor, or the like, and performs various control processing, arithmetic processing, and the like by reading and executing a program P (program product) and a database stored in advance in the storage unit 12. Note that a database server or the like may be provided outside the information processing device 1, and the control unit 11 may read the database from the database server or the like. In addition, the information processing device 1 may be one whose functions are realized by a plurality of server devices or computers. In addition, the information processing device 1 may correspond to a node on a blockchain.

[0014] The storage unit 12 of the information processing device 1 is, for example, a volatile memory and a non-volatile memory. The storage unit 12 stores a program P, a learning model M, a diagnostic information table 121, and a classification information history table 122. The program P may be provided to the information processing device 1 using a storage medium 12a in which the program P is stored so as to be readable by a computer. The storage medium 12a is, for example, a portable memory. Examples of the portable memory include a CD-ROM, a USB (Universal Serial Bus) memory, an SD card, a micro SD card, and a flash memory. When the storage medium 12a is a portable memory, the processing element of the control unit 11 may read the program P from the storage medium 12a using a reading device (not shown). The read program P is written to the storage unit 12. Furthermore, the program P may be provided to the information processing device 1 by the communication unit 13 communicating with an external device. When answer information is input, the learning model M outputs classification information indicating the type of headache suffered by the user. The diagnostic information table 121 stores information (diagnosis information) related to headache diagnosis that corresponds to the classification information output by the learning model M and that is transmitted by the information processing device 1 to the terminal device 2. The diagnostic information table 121 stores a history of classification information that has been output in the past by the learning model M. Details of the learning model M, the diagnostic information table 121, and the classification information history table 122 will be described later.

[0015] The communication unit 13 is a communication module or a communication interface for communicating with the terminal device 2 by wire or wirelessly, and is, for example, a wide-area wireless communication module such as LTE (registered trademark), 4G, or 5G. The control unit 11 communicates with the terminal device 2 via the communication unit 13 through an external network N such as the Internet.

[0016] FIG. 3 is a block diagram showing a configuration example of the terminal device 2. The terminal device 2 is, for example, a smartphone, a tablet terminal, wearable glasses, a wearable computer, or a personal computer. In this embodiment, an example in which the terminal device 2 is a smartphone will be described. The terminal device 2 includes a terminal control unit 21, a storage unit 22, a communication unit 23, a display unit 24, and an input unit 25. The terminal control unit 21 is configured with a CPU or an MPU, and performs various control processes, arithmetic processes, and the like. The functions of the terminal device 2 may be realized by multiple devices. The terminal device 2 may also execute a part or all of the processes performed by the information processing device 1.

[0017] The storage unit 22 stores an application program (app program) Pa that receives answer information input from a user, transmits the input answer information to the information processing device 1, and displays classification information (diagnosis result) and diagnosis information received from the information processing device 1. The app program Pa is provided to the terminal device 2 using, for example, a storage medium 22a. Note that the terminal control unit 21 of the terminal device 2 may obtain the app program Pa using the Internet and store it in the storage unit 22.

[0018] The communication unit 23 is a communication module or a communication interface for wirelessly communicating with the information processing device 1. The terminal control unit 21 communicates with the information processing device 1 through the communication unit 23 and an external network N.

[0019] The display unit 24 displays an answer information input screen for receiving input of answer information, a diagnosis result display screen for displaying the classification information (diagnosis result) received from the information processing device 1, and the like.

[0020] The input unit 25 accepts input of the answer information. In this embodiment, the terminal device 2 is a smartphone, and the display unit 24 and the input unit 25 are integrally configured with a touch panel. The terminal device 2 may acquire the answer information by photographing or scanning a paper questionnaire.

[0021] FIG. 4 is an explanatory diagram showing an example of the learning model M according to the first embodiment. When answer information is input, the learning model M outputs classification information indicating whether the headache suffered by the user is migraine, medication-overuse headache, tension-type headache, trigeminal autonomic headache, other primary headache, or secondary headache. The classification information may indicate either primary headache or secondary headache. Migraine, medication-overuse headache, tension-type headache, trigeminal autonomic headache, and other primary headaches are included in primary headaches. The answer information according to this embodiment includes answer data input by the user for each item including age, sex, height, weight, headache onset age, headache frequency, headache duration, headache site, headache characteristics, headache severity, headache improvement or worsening degree due to exercise (exercise effect), accompanying symptoms, aura, headache time period, headache trigger, medication information (type of medicine, frequency of taking, and efficacy), and family headache information. The answer information may include some of the above items. The answer information may also include answer data related to other items such as medical history or lifestyle.

[0022] The learning model M according to this embodiment is configured by a decision tree. The learning model M may be a neural network such as Transformer, BERT (Bidirectional Encoder Representations from Transformers), GPT (Generative Pre-trained Transformer), CNN (Convolution Neural Network), RNN (Recurrent Neural Network), or LSTM (Long Short Term Memory), or may be other learning algorithms such as support vector machine, logistics regression, random forest, or XGBoost (eXtreme Gradient Boosting). When input data (answer information) is input to the learning model M, the input data (answer information) is classified according to conditions on the way from the root to the branch, and when it reaches the terminal leaf node, the value given to the terminal leaf node is output as classification information (diagnosis result). That is, the output data of the learning model M is either migraine, medication-overuse headache, tension-type headache, trigeminal autonomic headache, other primary headache, or secondary headache. The output data of the learning model M may include classification information indicating no disease (normal). The learning model M may also output the likelihood that each headache type matches the headache type suffered by the user, and the headache type with the highest likelihood may be the diagnosis result.

[0023] The learning model M is trained using training data that links answer information with the correct answer value label of the classification information to be output. The correct answer value label is based on, for example, a result of a diagnosis made by a headache specialist based on the answer information. The learning model M is trained using, for example, training data stored in the storage unit 12 of the information processing device 1. The learning model M is trained by, for example, changing the order of the conditions of each node so that, when answer information included in the training data is input, the learning model M outputs a value equal to the correct answer value label. For example, when the diagnosis result by the specialist based on the answer information is migraine, the learning model M is trained to output migraine when the answer information is input. The learning model M is trained (constructed) by performing the same process on a large number of training data. The learning model M may be trained in a device other than the information processing device 1.

[0024] The control unit 11 of the information processing device 1 calculates the contribution of each item included in the answer information input to the learning model M to the output classification information. The method of calculating the contribution is not limited, but for example, SHAP (SHapley Additive exPlanation), LIME (Local Interpretable Model-Agnostic Explanations), Attention mechanism, etc. can be used. The greater the contribution of an item, the greater the influence it has on the output of the classification information. When the contribution is calculated by SHAP, the contribution is a value between 0 and 1. The control unit 11 transmits the item with the highest contribution among the items included in the answer information to the terminal device 2.

[0025] FIG. 5 is an explanatory diagram showing an example of the diagnostic information table 121. The diagnostic information table stores information on headache diagnosis (name of treatment hospital and name of treatment drug presented to the user) corresponding to classification information (diagnosis result). The management items (fields) of the diagnostic information table 121 include a classification information field, a treatment hospital field, and a treatment drug field. The classification information field stores classification information (type of headache) of the headache suffered by the user, which is output by the learning model M. The treatment hospital field stores the name of the treatment hospital that can treat the headache, which is presented to the user for the classification information. The treatment drug field stores the name of the treatment drug required for the treatment of the headache type. The diagnostic information table 121 may include a treatment drug amount field in which the amount of the treatment drug required for treatment (frequency or dose per time) is stored. When the terminal device 2 is a terminal owned by the user, the treatment drug table stores the names of treatment drugs, which are prescription drugs or over-the-counter drugs. When the terminal device 2 is a terminal owned by a hospital or a doctor who examines headache patients, the treatment drug table stores the names of treatment drugs, which are prescription drugs. The control unit 11 of the information processing device 1 refers to the diagnostic information table 121 and identifies the name of the treating hospital and the name of the therapeutic drug corresponding to the classification information. The control unit 11 transmits the identified name of the treating hospital and the name of the therapeutic drug together with the classification information to the terminal device 2. The diagnostic information table 121 may include a region field in which the region in which the treating hospital is located is stored. In this case, the control unit 11 may identify the name of the treating hospital to be presented to the user based on the classification information, the region in which the user lives, and the region in which the hospital is located.

[0026] FIG. 6 is an explanatory diagram showing an example of the classified information history table 122. The classified information history table 122 stores a history of classified information previously output to a user. The management items (fields) of the classified information history table include a date field and a classified information field. The date (year, month, and day) of the day the classified information was output is stored in the date field, and the output classified information is stored in the classified information field. The control unit 11 of the information processing device 1 adds a record linking the output date and the output classified information to the classified information history table 122 every time the learning model M outputs classified information. In addition, when transmitting classified information output by the learning model M to the terminal device 2, the control unit 11 reads, for example, the most recent five pieces of classified information and the dates on which the classified information was output from the classified information history table 122, and also transmits the read classified information and the dates on which the classified information was output to the terminal device 2. The storage unit of the information processing device 1 stores the classified information history table 122 for each of a plurality of users. The classified information history table 122 may be stored in the terminal device 2 owned by each user.

[0027] 7 and 8 are explanatory diagrams showing an example of a medical questionnaire screen in the terminal device 2. The terminal control unit 21 of the terminal device 2 causes the display unit 24 to display an input field for each item of the answer information, and causes the input unit 25 to accept input of answer data (answer information) for each item. The input unit 25 accepts input of age, sex, height, weight, headache onset age, headache frequency, headache duration, headache site, headache characteristics, headache severity, headache improvement or worsening degree due to exercise, accompanying symptoms, aura, headache time period, headache trigger, medication information (type of medication being taken, frequency of taking, and effectiveness), and family headache information by selection on the medical questionnaire screen. The user can input answer data by tapping a check box provided for each option for each item. In this embodiment, the input unit 25 accepts selection of one option for each item, but may accept selection of multiple options. The input unit 25 may also accept input of answer information by writing in instead of selecting. In this case, the written answer information is input to a learning model M capable of natural language processing.

[0028] 9 is an explanatory diagram showing an example of a diagnosis result display screen in the terminal device 2. The terminal control unit 21 of the terminal device 2 causes the display unit 24 to display the diagnosis result (category information) received from the information processing device 1, the answer information item with the highest contribution and answer data, the name of the treating hospital and the name of the therapeutic drug (information related to headache diagnosis), and the history of the category information output in the past (the most recent five pieces of category information and the dates when the category information was output).

[0029] 10 is a flowchart showing an example of processing of the information processing device 1 and the terminal device 2 according to the first embodiment. The terminal control unit 21 of the terminal device 2 causes the display unit 24 to display a medical questionnaire screen (S1). The terminal control unit 21 causes the input unit 25 to accept input of answer information from the user (S2). The terminal control unit 21 transmits the input answer information to the information processing device 1 (S3).

[0030] The control unit 11 of the information processing device 1 receives the answer information from the terminal device 2 (S4). The control unit 11 inputs the received answer information to the learning model M (S5) and outputs the classification information (S6). The control unit 11 also calculates the contribution of each item of the answer information input in S5 to the output of the classification information (S7). The control unit 11 identifies the item with the highest contribution among the items included in the answer information (S8). The control unit 11 refers to the diagnosis information table 121 and identifies the name of the treating hospital and the name of the medicine (information related to headache diagnosis) based on the classification information output in S6 (S9). The control unit 11 links the output classification information with the date on which the classification information was output, and adds it to a new record in the classification information history table 122 and stores it (S10). The control unit 11 reads out the last five pieces of classification information and the dates on which the classification information was output from the classification information history table 122 (S11). The control unit 11 transmits to the terminal device 2 (S12) the classification information output in S6, the item identified in S8 (the item with the highest contribution), and the last five pieces of classification information and the dates on which the classification information was output read in S10, and terminates the processing.

[0031] The terminal control unit 21 of the terminal device 2 receives the classification information, the item with the highest contribution rate, and the most recent five pieces of classification information and the dates on which the classification information was output from the information processing device 1 (S13). The terminal control unit 21 causes the display unit 24 to display the classification information, the item with the highest contribution rate, the name of the treating hospital, the name of the therapeutic drug, and the most recent five pieces of classification information and the dates on which the classification information was output (S14), and ends the process.

[0032] According to the above configuration and processing, it is possible to easily determine the type of headache based on the answer information input by the user. Note that the input data of the learning model M may include the classification information stored in the classification information history table 122 and output by the learning model M in the past.

[0033] (Embodiment 2) The control unit 11 of the information processing device 1 according to the second embodiment stores the answer information input by the user to the terminal device 2 and transmitted to the information processing device 1 in the memory unit 12 as headache history data. The control unit 11 inputs the previously stored time-series answer information (headache history data) into the learning model M, and outputs classification information and information regarding therapeutic drugs required for treatment. The following describes the differences between the second embodiment and the first embodiment. The other configurations, except for those described below, are the same as those of the first embodiment. For this reason, the components common to the first embodiment are given the same reference symbols as those of the first embodiment, and descriptions of those components will be omitted.

[0034] 11 is a block diagram showing a configuration example of an information processing device 1 according to embodiment 2. The storage unit 12 of the information processing device 1 according to embodiment 2 stores a headache history table 123. In addition, the learning model M according to embodiment 2 includes a first learning model M1 and a second learning model M2.

[0035] FIG. 12 is an explanatory diagram showing an example of the headache history table 123. The headache history table 123 stores answer information (headache history data) that was previously input by a user to the terminal device 2 and transmitted to the information processing device 1. The management items of the headache history table 123 include, for example, a date field, a headache frequency field, a headache duration field, a headache location field, a headache characteristics field, a medication taken field, a medication frequency field, and an efficacy field. The date field stores the date (year, month, and day) on which the answer information was input to the terminal device 2. The headache frequency field, headache duration field, headache location field, and headache characteristics field store the headache frequency, headache duration, headache location, and headache characteristics included in the input information. The medication taken field, medication frequency field, and efficacy field store medication information (type of medication being taken, medication frequency, and efficacy) included in the answer information. When the control unit 11 of the information processing device 1 receives input answer information from the terminal device 2, it adds the headache frequency, headache duration, headache location, headache characteristics, and medication information included in the answer information to a new record in the headache history table 123. Note that the headache history table 123 may store items included in the answer information other than the items described above.

[0036] FIG. 13 is an explanatory diagram showing an example of the learning model M according to the second embodiment. The learning model M according to the present embodiment includes a first learning model M1 and a second learning model M2. Answer information (time-series answer information) related to a plurality of records stored in the headache history table 123 is input to the first learning model M1 and the second learning model M2. The control unit 11 of the information processing device 1 reads, for example, records of answer information for the most recent five times from the headache history table 123, and inputs the read answer information to the first learning model M1 and the second learning model M2 as headache history data (input data).

[0037] The first learning model M1 is configured by, for example, Transformer. The first learning model M1 may be, for example, another neural network such as BERT, GPT, RNN, CNN, or LSTM, or another learning algorithm such as a decision tree, a support vector machine, a logistics regression, a random forest, or XGBoost. The input layer of the first learning model M1 accepts the input of answer information for each date included in the headache history data in date order. The intermediate layer of the second learning model M2 has a plurality of neurons that extract features of the input data, and transfers the extracted features to the output layer. The output layer has neurons that output the name of a therapeutic drug required for the treatment of the user, and outputs classification information (diagnosis result) based on the features output from the intermediate layer. When the first learning model M1 is a Transformer, the control unit 11 of the information processing device 1 may derive the contribution degree of each item to the output of the classification information using the attention mechanism of the Transformer. That is, the control unit 11 derives how much attention the first learning model M1 has shown to each item (token) as the contribution degree.

[0038] The second learning model M2 is configured by, for example, Transformer. The second learning model M2 may be, for example, other neural networks such as BERT, GPT, RNN, CNN, or LSTM, or other learning algorithms such as decision tree, support vector machine, logistics regression, random forest, or XGBoost. The input layer of the second learning model M2 accepts the input of answer information for each date included in the headache history data in date order. The intermediate layer of the second learning model M2 has a plurality of neurons that extract features of the input data, and the output layer that passes the extracted features to the output layer has neurons that output information about the therapeutic drug necessary for the user's treatment, and outputs information about the therapeutic drug necessary for the user's treatment (the name of the therapeutic drug) based on the features output from the intermediate layer. When the second learning model M2 is a Transformer, the control unit 11 of the information processing device 1 may derive the contribution degree of each item to the output of information about the therapeutic drug using the attention mechanism of the Transformer. That is, the control unit 11 derives how much attention the second learning model M2 has shown to each item (token) as the contribution degree.

[0039] As described above, when headache history data including time-series response information is input, the learning model M according to embodiment 2 outputs classification information (diagnosis result) and information on the therapeutic drug required for the user's treatment (therapeutic drug name). Note that the learning model M may output only either the classification information or the therapeutic drug name. When the terminal device 2 is a terminal owned by the user, the therapeutic drug name output by the learning model M is the therapeutic drug name of a prescription drug or an over-the-counter drug. When the terminal device 2 is a terminal owned by a hospital or a doctor who examines headache patients, the therapeutic drug name output by the learning model M is the therapeutic drug name of a prescription drug.

[0040] The first learning model M1 is trained using training data that links headache history data with a correct value label of the classification information to be output. The correct value label is based on, for example, a diagnosis made by a headache specialist based on headache history data. The first learning model M1 is trained by correcting the weights and biases using error backpropagation so that, when headache history data is input, the first learning model M1 outputs a value equal to the correct value label. For example, if the diagnosis made by a specialist based on headache history data is migraine, the learning model M is trained to output migraine when the headache history data is input. The learning model M is trained (constructed) by performing similar processing on a large number of training data.

[0041] The second learning model M2 is trained using training data that links headache history data with a correct value label of the name of the therapeutic drug to be output. The correct value label is, for example, the name of a therapeutic drug that was actually prescribed to a patient by a headache specialist and that alleviated or cured the patient's headache (was effective). The second learning model M2 is trained by correcting the weights and biases using backpropagation so that, when headache history data is input, the model outputs the same value as the correct value label. For example, if the name of a therapeutic drug that was prescribed based on headache history data and was effective is drug U, the second learning model M2 is trained to output drug U when the headache history data is input. The learning model M is trained (constructed) by performing similar processing on a large number of training data.

[0042] 14 is a flowchart showing an example of processing of the information processing device 1 and the terminal device 2 according to the second embodiment. The terminal control unit 21 of the terminal device 2 causes the display unit 24 to display a medical questionnaire screen (S21). The terminal control unit 21 causes the input unit 25 to accept input of answer information from the user (S22). The terminal control unit 21 transmits the input answer information to the information processing device 1 (S23).

[0043] The control unit 11 of the information processing device 1 receives the answer information from the terminal device 2 (S24). The control unit 11 stores the headache frequency, headache duration, headache location, headache characteristics, and medication information (type of medication being taken, frequency of taking, and effectiveness) from the received answer information in the headache history table 123 of the storage unit 12 (S25). The control unit 11 reads out the time-series answer information (headache history data) for the most recent five episodes from the headache history table 123 (S26). The control unit 11 inputs the read headache history data into the learning model M (S27), and outputs the classification information and the name of the therapeutic drug required for the user's treatment (S28). The control unit 11 transmits the output classification information and the name of the therapeutic drug required for the treatment to the terminal device 2 (S29), and ends the process.

[0044] The terminal control unit 21 of the terminal device 2 receives the classification information and the therapeutic drug name from the information processing device 1 (S30). The terminal control unit 21 causes the display unit 24 to display the classification information and the therapeutic drug name (S31), and ends the process.

[0045] The input data for the learning model M (the first learning model M1 and the second learning model M2) may include classification information or information on drugs output by the learning model M in the past. The learning model M may be configured as a single model that outputs both classification information and information on drugs. When the learning model M is configured as a model capable of outputting a time series such as GPT or seq2seq, the learning model M may output a prediction of classification information of the headache that the user will suffer from in the future, or a prediction of the name of the therapeutic drug and the amount of the therapeutic drug required in the future (information on the therapeutic drug) in a time series. In this case, the learning model M can output a prediction that the headache will be cured (return to normal) in the future, or a prediction that the therapeutic drug will not be required in the future. For example, the learning model M outputs classification information by time series such as migraine now, migraine in one month, and normal in three months. Also, the learning model M outputs information on the therapeutic drug by time series such as drug U three times a day now, drug U once a day in one month, and no medication required in three months. In this case, the learning model M is trained using training data including past time-series response data (headache history data) and classification information over multiple periods, and training data including past time-series response data (headache history data) and the names and dosages of medications prescribed to patients by headache specialists or effective medications over multiple periods.

[0046] (Embodiment 3) The terminal device 2 according to the third embodiment stores an application program for recording information about headaches, including headache frequency, headache duration, headache location, or headache characteristics, and an application program for recording information about medications being taken. The control unit 11 of the information processing device 1 acquires records about headaches or records about medications being taken from the application program executed by the terminal device 2, and outputs classification information or information about necessary medications (medicinal drug names) by inputting the acquired headache history data or records about medications being taken as answer information into the learning model. The following describes the differences between the third embodiment and the first and second embodiments. The other configurations, except for the configurations described below, are common to the first and second embodiments. For this reason, the same reference symbols as those in the first embodiment are used for components common to the first and second embodiments, and descriptions of those components are omitted.

[0047] FIG. 15 is a block diagram showing an example of the configuration of the terminal device 2 according to the third embodiment. The terminal device 2 according to the third embodiment stores a headache record application program (headache record app program) Pb, a medication record application program (medication record app program) Pc, a headache record table 221, and a medication record table 222. The headache record table 221 is read by the headache record app program Pb and displayed on the display unit 24. The medication record table 222 is read by the medication record app program Pc and displayed on the display unit 24. The headache record app program Pb and the medication record app program Pc are provided to the terminal device 2 using, for example, a storage medium 22a. The terminal control unit 21 of the terminal device 2 may obtain the headache record app program Pb and the medication record app program Pc using the Internet and store them in the storage unit 22. The storage unit 22 may store electronic medical record data obtained (downloaded) from the information processing device 1 or another device. In this case, the terminal control unit 21 transmits the information read out from the electronic medical record data to the information processing device 1 as response information.

[0048] FIG. 16 is an explanatory diagram showing an example of the headache record table 221. The headache record table 221 stores records related to headaches previously input by the user. The management items (fields) of the headache record table 221 include, for example, a date field, a headache frequency field, a headache duration field, a headache location field, and a headache characteristics field. The date (year, month, day) on which the user inputs answer information to the terminal device 2 is stored in the date field. The headache frequency field, headache duration field, headache location field, and headache characteristics field store the headache frequency, headache duration, headache location, and headache characteristics included in the answer information input by the user.

[0049] FIG. 17 is an explanatory diagram showing an example of the medication record table 222. The medication record table 222 stores records of medications currently being taken that were previously input by the user. The management items (fields) of the medication record table 222 include a date field, a medication taken field, a medication frequency field, and an efficacy field. The date (year, month, and day) on which the user inputs answer information to the terminal device 2 is stored in the date field. The medication taken field, medication frequency field, and efficacy field store medication information (type of medication currently being taken, medication frequency, and efficacy) included in the answer information input by the user.

[0050] FIG. 18 is an explanatory diagram showing an example of a headache record selection screen. When the terminal control unit 21 of the terminal device 2 starts executing the headache record application program Pb, it reads out the headache record table 221 from the storage unit 22 and displays it on the display unit 24. The user can input the selection of the headache record record to be transmitted to the information processing device 1 as answer information by the input unit 25. In the headache record selection screen according to this embodiment, multiple headache record records are selected. The terminal control unit 21 transmits the multiple selected headache records to the information processing device 1. That is, multiple pieces of answer information in time series (headache frequency, headache duration, headache site, and headache characteristics) are transmitted to the information processing device 1 as headache history data. In the example shown in FIG. 18, the selected headache record record is indicated by a black triangle. Note that the headache record table 221 according to this embodiment includes a headache frequency field, and the recorded headache frequency is included in the answer information, but is not limited to this. The terminal control unit 21 may record the headache record entered by the user each time the user develops a headache in the headache record table 221, linking it to the date entered, and include the headache frequency calculated based on multiple such dates in the response information or headache history data.

[0051] FIG. 19 is an explanatory diagram showing an example of a taken medicine record selection screen. When the terminal control unit 21 of the terminal device 2 starts executing the medication record application program Pc, it reads out the medication record table 222 from the storage unit 22 and displays it on the display unit 24. The user can input the selection of the record of the record related to the medicine to be transmitted to the information processing device 1 as the answer information by the input unit 25. In the taken medicine record selection screen according to this embodiment, the record of the record related to a plurality of medicines is selected. The terminal control unit 21 transmits the records related to the plurality of selected medicines to the information processing device 1. That is, a plurality of pieces of answer information (type of medicine being taken, frequency of taking, and effectiveness) in chronological order are transmitted to the information processing device 1. In the example shown in FIG. 19, the record of the record related to the selected medicine is indicated by a black triangle.

[0052] In this embodiment, a record of the medication record having the same date as the headache record record selected on the headache record selection screen shown in Fig. 18 is selected on the medication record selection screen shown in Fig. 19, but this is not limited to this. The date of the headache record record and the date of the record on the medication record selection screen may be different. Furthermore, the terminal control unit 21 of the terminal device 2 may automatically read out the records of the most recent five headache records and send them to the information processing device 1 when it starts executing the headache record application program Pb, and may automatically read out the records of the most recent five medication records and send them to the information processing device 1 when it starts executing the medication record application program Pc.

[0053] Fig. 20 is a flowchart showing an example of the processing of the information processing device and the terminal device according to the third embodiment. The terminal control unit 21 of the terminal device 2 causes the display unit 24 to display a headache record selection screen (S41). The terminal control unit 21 causes the input unit 25 to accept the selection of a record of a plurality of headache records (S42). The terminal control unit 21 causes the input unit 25 to display a medication record screen (S43). The terminal control unit 21 causes the input unit 25 to accept the selection of a record of a plurality of medications (S44). The terminal control unit 21 transmits the selected headache records and records of a plurality of medications to the information processing device 1 as headache history data including chronological response information (S45).

[0054] The control unit 11 of the information processing device 1 receives the headache history data from the terminal device 2 (S46). The control unit 11 inputs the received headache history data into the learning model M (S47) and outputs the classification information and the name of the treatment drug required for the user's treatment (S48). The control unit 11 transmits the output classification information and treatment name to the terminal device 2 (S49) and ends the process.

[0055] The terminal control unit 21 of the terminal device 2 receives the classification information and the therapeutic drug name from the information processing device 1 (S50). The terminal control unit 21 causes the display unit 24 to display the classification information and the therapeutic drug name (S51), and ends the process.

[0056] (Embodiment 4) The information processing device 1 according to the fourth embodiment stores the contribution degree of each item included in the answer information input to the calculated learning model M to the output classification information, and calculates the average value of the contribution degree of a plurality of answer information for each item. The information processing device 1 creates a graph by accumulating the average value of the contribution degree for each classification information item by answer information item, and transmits (outputs) it to the terminal device 2. The terminal device 2 according to the second embodiment is a terminal owned by a hospital or a doctor who examines headache patients, and displays a graph showing the average value of the contribution degree received from the information processing device 1. The following describes the differences between the fourth embodiment and the first embodiment. The other configurations, except for the configuration described later, are common to the first embodiment. For this reason, the same reference symbols as those in the first embodiment are given to the components common to the first embodiment, and the description of those components is omitted.

[0057] 21 is a block diagram showing an example of the configuration of the information processing device 1 according to embodiment 4. The storage unit 12 of the information processing device 1 according to embodiment 4 stores a contribution degree table .

[0058] FIG. 22 is an explanatory diagram showing an example of the contribution degree table 124. In the contribution degree table 124, the contribution degree of each item included in the answer information input to the learning model M to the output classification information is stored. The management items (fields) of the contribution degree table 124 include a classification information field and a plurality of contribution degree fields. Note that the contribution degree field includes a field in which the contribution degree related to each item included in the answer information is stored, but in FIG. 22, the contribution degree fields related to age, sex, height, and weight are shown, and the contribution degree fields related to other items are omitted. After outputting the classification information by the learning model M, the control unit 11 of the information processing device 1 calculates the contribution degree of each item to the output classification information, links it to the classification information, adds it to a new record in the contribution degree table, and stores it.

[0059] 23 is an explanatory diagram showing an example of a contribution degree graph display screen. For example, when a predetermined number of records are added to the contribution degree table, the control unit 11 of the information processing device 1 creates a graph (contribution degree graph) showing the average value of the contribution degree of each item of answer information for each piece of classification information, and transmits it to the terminal device 2. The terminal control unit 21 of the terminal device 2 causes the display unit 24 to display the contribution degree graph received from the information processing device 1. Note that, when information instructing the creation of a contribution degree graph is transmitted from the terminal device 2, the control unit 11 of the information processing device 1 may create the contribution degree graph and transmit it to the terminal device 2.

[0060] When a predetermined number (for example, 100) of records are added to the contribution table 124, the control unit 11 of the information processing device 1 calculates the average value of the contribution degree of each item for each classification information. In this embodiment, it is assumed that migraine and medication-overuse headache are the same classification information, and the average value of the contribution degree of each item is calculated for the records whose classification information is migraine and the records whose classification information is medication-overuse headache together. Specifically, the control unit 11 reads out all records whose classification information is migraine or medication-overuse headache, and calculates the average value of the contribution degree of each item for the read records. The control unit 11 similarly calculates the average value of the contribution degree of each item for the records whose classification information is medication-overuse headache, tension-type headache, trigeminal autonomic headache, other primary headaches, and secondary headaches. The control unit 11 creates a contribution degree graph in the form of a stacked graph by stacking up the average value of the contribution degree to the output of each classification information for each item of answer information. That is, the control unit 11 creates a graph by stacking up the average value of the contribution degree for each classification information for each item of answer information. The vertical axis of the contribution degree graph indicates the item of answer information, and the horizontal axis indicates the contribution degree. The contribution degree according to this embodiment is calculated by SHAP, and the value of the contribution degree for each classification information is in the range of 0 to 1. In the example shown in FIG. 23, the answer information items are arranged from the top in descending order of the total contribution degree to the output of each classification information. Also, in this embodiment, the average contribution degree is calculated for each answer information item, but the average contribution degree may be calculated for each answer data (option) related to the answer information item.

[0061] The embodiments disclosed herein should be considered to be illustrative in all respects and not restrictive. The technical features described in each embodiment can be combined with each other, and the scope of the present invention is intended to include all modifications within the scope of the claims and equivalents to the scope of the claims. In addition, the independent claims and dependent claims described in the claims can be combined with each other in all combinations regardless of the reference format. Furthermore, the claims use a format in which a claim references two or more other claims (multiple claim format), but this is not limited to this. A format in which a multiple claim (multi-multi claim) that references at least one multiple claim may be used. [Explanation of symbols]

[0062] 1. Information processing device 11 Control section 12 Storage section 121 Diagnostic Information Table 122 Taxonomic Information History Table 123 Headache History Table 124 Contribution Table 12a Storage medium 13. Communications Department 2 Terminal Equipment 21 Terminal control unit 22 Memory section 221 Headache Record Table 222 Medication Record Table 22a Storage medium 23 Communications Department 24 Display section 25 Input section M Learning Model M1 First learning model M2 Second Learning Model N Network P Program Pa Application Program (Application Program) Pb Headache Record Application Program (Headache Record App Program) Pc medication record application program (medication record app program) S Information Processing System

Claims

1. Acquire answer information including multiple items of a user regarding headaches, inputting the acquired answer information into a learning model trained to output classification information indicating either a primary headache including migraine or a secondary headache when answer information is input, thereby outputting the classification information; outputting a contribution of the item of the answer information to the output of the classification information; outputting the item having the highest contribution; outputting information related to headache diagnosis, including the names of hospitals that can treat headaches and the names of therapeutic drugs required for the treatment of the type of headache, corresponding to the classification information; outputting a history of the classification information that has been output in the past; The classification information, the items with high contribution rates, information regarding the headache diagnosis, and the history of the classification information are displayed on the same screen. A program that causes a computer to carry out processing.

2. Acquire answer information including multiple items of a user regarding headaches, inputting the acquired answer information into a learning model trained to output classification information indicating either a primary headache including migraine or a secondary headache when answer information is input, thereby outputting the classification information; outputting a contribution of the item of the answer information to the output of the classification information; storing the contribution of each of the items to the output classification information; Calculate the average contribution of each of the items for each classification information; A graph is output in which the vertical axis indicates the item of the answer information, and the horizontal axis indicates a total obtained by accumulating the average values ​​of the contribution rates for each of the classified information by the item, and the display mode of each part showing the average value of each of the classified information is different. A program that causes a computer to carry out processing.

3. Outputting the classification information indicating either migraine, medication overuse headache, tension-type headache, trigeminal autonomic headache, other primary headache, or secondary headache.

3. The program according to claim 1 or 2.

4. The answer information includes items such as headache frequency, headache duration, headache location, headache characteristics, or information about medications being taken.

3. The program according to claim 1 or 2.

5. Outputting the contribution of the item of the answer information to the output of the classification information The program according to claim 4.

6. Outputting information regarding headache diagnosis corresponding to the classification information 3. The program according to claim 1 or 2.

7. Output the history of the classification information that has been output in the past 3. The program according to claim 1 or 2.

8. Obtain headache history data including the time-series answer information; The learning model is trained to output the classification information when headache history data is input, The acquired headache history data is input to the learning model, and the classification information is output.

3. The program according to claim 1 or 2.

9. Obtain headache history data including the time-series answer information; The learning model is trained to output information about a necessary therapeutic drug when headache history data is input, The acquired headache history data is input into the learning model, and information regarding the necessary therapeutic drug is output.

3. The program according to claim 1 or 2.

10. obtaining a headache record or a medication record from an application program that provides a headache record or a medication record, the headache record including headache frequency, headache duration, headache location, or headache characteristics; The acquired headache record or the record of the medication being taken is input as answer information into the learning model, and the classification information is output.

3. The program according to claim 1 or 2.

11. Obtaining the record of headache history data or the record of medication being taken from an application program that records the headache history data or the record of medication being taken; The headache history data or the record of the medication being taken is input as the answer information into the learning model, and the classification information is output. The program according to claim 8.

12. Obtaining the record of headache history data or the record of medication being taken from an application program that records the headache history data or the record of medication being taken; The acquired headache history data or records regarding medication being taken are input into the learning model as the answer information, and information regarding the necessary medication is output. The program according to claim 9.

13. A graph is output by accumulating the average value of the contribution degree for each of the classification information items. The program according to claim 5.

14. Acquire answer information including multiple items of a user regarding headaches, inputting the acquired answer information into a learning model trained to output classification information indicating either a primary headache including migraine or a secondary headache when answer information is input, thereby outputting the classification information; outputting a contribution of the item of the answer information to the output of the classification information; outputting the item having the highest contribution; outputting information related to headache diagnosis, including the names of hospitals that can treat headaches and the names of therapeutic drugs required for the treatment of the type of headache, corresponding to the classification information; outputting a history of the classification information that has been output in the past; The classification information, the items with high contribution rates, information regarding the headache diagnosis, and the history of the classification information are displayed on the same screen. Information processing methods.

15. Acquire answer information including multiple items of a user regarding headaches, inputting the acquired answer information into a learning model trained to output classification information indicating either a primary headache including migraine or a secondary headache when answer information is input, thereby outputting the classification information; outputting a contribution of the item of the answer information to the output of the classification information; storing the contribution of each of the items to the output classification information; Calculate the average contribution of each of the items for each classification information; A graph is output in which the vertical axis indicates the item of the answer information, and the horizontal axis indicates a total obtained by accumulating the average values ​​of the contribution rates for each of the classified information by the item, and the display mode of each part showing the average value of each of the classified information is different. Information processing methods.

16. Acquire answer information including multiple items of a user regarding headaches, inputting the acquired answer information into a learning model trained to output classification information indicating either a primary headache including migraine or a secondary headache when answer information is input, thereby outputting the classification information; outputting a contribution of the item of the answer information to the output of the classification information; outputting the item having the highest contribution; outputting information related to headache diagnosis, including the names of hospitals that can treat headaches and the names of therapeutic drugs required for the treatment of the type of headache, corresponding to the classification information; outputting a history of the classification information that has been output in the past; The classification information, the items with high contribution rates, information regarding the headache diagnosis, and the history of the classification information are displayed on the same screen. Control Unit An information processing device comprising:

17. Acquire answer information including multiple items of a user regarding headaches, inputting the acquired answer information into a learning model trained to output classification information indicating either a primary headache including migraine or a secondary headache when answer information is input, thereby outputting the classification information; outputting a contribution of the item of the answer information to the output of the classification information; storing the contribution of each of the items to the output classification information; Calculate the average contribution of each of the items for each classification information; A graph is output in which the vertical axis indicates the item of the answer information, and the horizontal axis indicates a total obtained by accumulating the average values ​​of the contribution rates for each of the classified information by the item, and the display mode of each part showing the average value of each of the classified information is different. Control Unit An information processing device comprising:

Images