Pulse wave information processing device, pulse wave information processing method, and pulse wave information processing program

Abstract

This information processing device comprises a prescription candidate output unit that, on the basis of a first time frequency spectrum image indicating a first time frequency spectrum corresponding to a waveform of a pulse wave of a first subject, outputs output information including Chinese medicine candidate information which indicates a candidate of a Chinese medicine to be prescribed to the first subject.

Description

[Technical Field]

[0001] The present disclosure provides: Pulse wave information processing device, Pulse wave Information Processing Method and pulse wave information processing Regarding the program. This application claims priority based on Japanese Patent Application No. 2022-074671, filed on April 28, 2022, the contents of which are incorporated herein by reference. [Background technology]

[0002] Research and development is underway into technologies that assist doctors in prescribing medications to patients.

[0003] In this regard, an information processing device is known that outputs information indicating a prescription for a Chinese herbal medicine associated with the symptoms of a subject whose symptoms are to be diagnosed by a Chinese herbal medicine doctor, based on a database containing information that associates, for each of a plurality of subjects, the symptoms that the subject has with the history of the prescription of the Chinese herbal medicine for the subject (see Patent Document 1). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2016-139268 Summary of the Invention [Problem to be solved by the invention]

[0005] Here, the symptoms of the subject described in Patent Document 1 include headache, dizziness, menopausal disorder, etc., and refer to the physical or mental state of the subject diagnosed by a traditional Chinese medicine doctor. For this reason, with an information processing device such as that described in Patent Document 1, the traditional Chinese medicine doctor must diagnose the symptoms based on diagnostic methods such as interviewing, facial examination, tongue examination, abdominal examination, and pulse examination, and input information indicating the diagnosis results into the information processing device. For this reason, there are cases where the information processing device cannot sufficiently reduce the effort required for the traditional Chinese medicine doctor to prescribe traditional Chinese medicine to the subject.

[0006] The present disclosure has been made in consideration of such circumstances, and can reduce the effort required for a Chinese medicine doctor to prescribe a Chinese medicine to a first subject. Pulse wave information processing device, Pulse wave Information Processing Method and pulse wave information processing The objective is to provide a program. [Means for solving the problem]

[0007] One aspect of the present disclosure is to measure the pulse wave of a first subject. a calculation unit that calculates a first time-frequency spectrum based on first waveform information that indicates the first waveform as a first waveform; First time-frequency spectrum a generation unit that generates a first time-frequency spectral image based on the a prescription candidate output unit that outputs output information including herbal medicine candidate information indicating candidates of herbal medicines to be prescribed to the first subject based on the first time-frequency spectral image; The prescription candidate output unit identifies candidates for herbal medicines to be prescribed to the first subject based on the first machine learning model, the second machine learning model, and the first time-frequency spectrum image, wherein the first machine learning model is a machine learning model that has been trained with first correspondence information in which a second time-frequency spectrum image showing a second time-frequency spectrum corresponding to a waveform of the pulse wave of the second subject is associated with second diagnosis result information showing a diagnosis result of the second subject by a herbal medicine doctor, and the second machine learning model is a machine learning model that has been trained with second correspondence information in which the second diagnosis result information is associated with herbal medicine information, and the herbal medicine information is information showing each of one or more herbal medicines prescribed by the herbal medicine doctor to the second subject. Ru, Pulse wave It is an information processing device.

[0008] Furthermore, one aspect of the present disclosure is In a computer, Pulse wave waveform of the first subject Based on first waveform information indicating the first waveform, The first time-frequency spectrum a calculation step of calculating the first time-frequency spectrum image based on the first time-frequency spectrum; and a generation step of generating a first time-frequency spectrum image based on the first time-frequency spectrum. a prescription candidate output step of outputting output information including herbal medicine candidate information indicating candidates of herbal medicines to be prescribed to the first subject based on the first time-frequency spectral image; and, With The prescription candidate output step identifies candidates for Chinese herbal medicines to be prescribed to the first subject based on the first machine learning model, the second machine learning model, and the first time-frequency spectrum image, wherein the first machine learning model is a machine learning model trained with first correspondence information in which a second time-frequency spectrum image showing a second time-frequency spectrum corresponding to a waveform of the pulse wave of the second subject is associated with second diagnosis result information showing a diagnosis result of the second subject by a Chinese herbal medicine doctor, and the second machine learning model is a machine learning model trained with second correspondence information in which the second diagnosis result information is associated with Chinese herbal medicine information, and the Chinese herbal medicine information is information showing each of one or more Chinese herbal medicines prescribed by the Chinese herbal medicine doctor to the second subject. Ru, Pulse wave It is an information processing method.

[0009] In one aspect of the present disclosure, a computer is configured to: Based on first waveform information indicating the first waveform, The first time-frequency spectrum a calculation step of calculating the first time-frequency spectrum image based on the first time-frequency spectrum; and a generation step of generating a first time-frequency spectrum image based on the first time-frequency spectrum. a prescription candidate output step of outputting output information including herbal medicine candidate information indicating candidates of herbal medicines to be prescribed to the first subject based on the first time-frequency spectral image. a pulse wave information processing program, wherein the prescription candidate output step identifies candidate herbal medicines to be prescribed to a first subject based on a first machine learning model, a second machine learning model, and a first time-frequency spectrum image, wherein the first machine learning model is a machine learning model that has learned first correspondence information in which a second time-frequency spectrum image showing a second time-frequency spectrum corresponding to a waveform of a pulse wave of the second subject is associated with second diagnosis result information showing a diagnosis result of the second subject by a Chinese medicine doctor, and the second machine learning model is a machine learning model that has learned second correspondence information in which the second diagnosis result information is associated with Chinese medicine information, and the Chinese medicine information is information showing each of one or more Chinese medicines prescribed by the Chinese medicine doctor to the second subject. It is a program. [Effects of the Invention]

[0010] According to the present disclosure, it is possible to reduce the effort required for a Chinese medicine doctor to prescribe a Chinese medicine to a first subject. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a diagram showing an example of the configuration of a pulse wave information processing system 1 including a pulse wave information processing device 20. [Figure 2] FIG. 2 is a diagram showing an example of the hardware configuration of a pulse wave information processing device 20. [Figure 3] FIG. 2 is a diagram showing an example of the functional configuration of a pulse wave information processing device 20. [Figure 4] FIG. 3 is a diagram showing an example of the flow of processing by the pulse wave information processing device 20 to generate waveform information. [Figure 5] FIG. 10 is a diagram showing an example of the flow of processing in which the pulse wave information processing device 20 causes the first machine learning model to learn the first correspondence information and causes the second machine learning model to learn the second correspondence information. [Figure 6] FIG. 10 is a diagram showing an example of a waveform indicated by learning waveform information associated with the target subject identification information selected in step S220. [Figure 7] FIG. 10 is a diagram showing an example of a time-frequency spectrum image. [Figure 8] FIG. 10 is a visualized image diagram illustrating an example of the processing in step S270. [Figure 9] FIG. 4 is a diagram showing an example of the flow of processing in which the pulse wave information processing device 20 receives diagnosis result information. [Figure 10] FIG. 10 is a diagram showing an example of an information reception image PCT1. [Figure 11] FIG. 10 is a diagram showing an example of how each of six drop-down menus is displayed. [Figure 12] FIG. 10 is a diagram showing an example of the flow of processing by which the pulse wave information processing device 20 outputs candidate herbal medicine information. [Figure 13] FIG. 10 is a diagram showing an example of the likelihood of each of a plurality of options included in a diagnostic item such as floating veins. [Figure 14] An image visualizing the process by which the second machine learning model identifies one or more candidate herbal medicines that are estimated to be plausible as the one or more herbal medicines to be prescribed to subject S3. [Figure 15] FIG. 10 is a diagram showing an example of an information reception image PCT2. [Figure 16] FIG. 10 is a diagram showing an example of the process by which a time-frequency spectrum image is generated by the pulse wave information processing device 20 based on waveform information indicating the waveform of a pulse wave detected by a pressure-varying arterial wave detection method. DETAILED DESCRIPTION OF THE INVENTION

[0012] <Embodiment> Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0013] < Pulse wave Overview of information processing equipment> First, according to this embodiment Pulse wave Information processing device (Hereinafter, simply referred to as "information processing device") This section provides an overview of the system.

[0014] Currently, Western medicine is mainstream due to its compatibility with science. However, there are other traditional medical practices around the world, such as Greek medicine, Arabic medicine (Unicine medicine), Indian medicine (Ayurveda), Tibetan medicine, Mongolian medicine, traditional Chinese medicine (Traditional Chinese Medicine), Korean medicine, and Wakan medicine (Kampo medicine). In Japan, the narrow definition of Oriental medicine, which utilizes acupuncture, moxibustion, medicinal foods, and herbal medicine derived from Traditional Chinese Medicine, is called Kampo medicine. This form of Kampo medicine has developed uniquely in Japan. In recent years, Japan has been reevaluating not only Western medicine but also complementary and alternative medicine. As a result, homeopathy, naturopathy, aromatherapy, and other therapies have also been gaining attention. In Japan, the combination of Western medicine and Kampo medicine as part of integrative medicine, including Kampo medicine, is becoming increasingly common, with approximately 90% of doctors prescribing Kampo medicine.

[0015] In Japan, health insurance is approved for several hundred types of herbal medicines, 148 types of medical herbal preparations (preparations approved by the Ministry of Health, Labor and Welfare), and a few decoctions.In addition to these herbal medicines approved for health insurance coverage, Japan also distributes over-the-counter (OTC) medicines (medicines that can be purchased without a prescription at pharmacies and drug stores) that are not covered by health insurance, and herbal medicines available for private treatment at specialized hospitals.

[0016] In addition to interviews, facial examinations, tongue examinations, and abdominal examinations, pulse diagnosis is also known as a diagnostic technique in traditional Chinese medicine. Pulse diagnosis is based on the idea that pulse waves reflect the state of an organ and involve diagnosing the characteristics of the pulse wave in relation to the organ's condition. For ease of explanation, the types of characteristics that appear in pulse waves depending on the organ's condition will be referred to simply as "pulse types." One method of classifying pulse types according to organ condition is the 28 Disease Pulses, which classifies the characteristics of pulse waves into 28 types depending on the organ's condition. Each of these 28 Disease Pulses is further divided into six pulse types: the Floating Net Pulse, the Shen Net Pulse, the Slow Net Pulse, the Su Net Pulse, the Xu Net Pulse, and the Shi Net Pulse. For ease of explanation, these six pulse types will be referred to as the Six Major Pulse Types. The Floating Net Pulse category includes the following six pulse types: the Bu Mai, the Kang Mai (a Chinese character with a grass crown over the hole), the Hong Mai, the Kang Mai, the Wei Mai, and the San Mai. The category of shenwang pulses includes four pulse types: shen, fu, xiang, and lao. The category of slowwang pulses includes five pulse types: slow pulse, brady, shibu, jie, and dai. The category of severalwang pulses includes three pulse types: several, arterial, and accelerating. The category of virtualwang pulses includes four pulse types: virtual, short, thin, and micro. The category of fullwang pulses includes six pulse types: real, long, xuan, tight, slippery, and large.

[0017] For example, a Chinese medicine doctor will diagnose the characteristics of a patient's pulse wave, identify the primary illness by assigning a diagnostic name to the eight primary pulses that frequently appear in the pulse wave out of 28 diseased pulses, and prescribe a type of Chinese medicine corresponding to the primary illness. In other words, a Chinese medicine doctor's pulse diagnosis is directly linked to the prescription of Chinese medicine. The effectiveness of Chinese medicine prescribed by a Chinese medicine doctor is based on statistics.

[0018] Information related to diagnosis by Chinese medicine doctors is explained in detail, for example, on the following reference websites, so further detailed explanation will be omitted in this specification. Reference website 1: Ministry of Health, Labour and Welfare website / Committee on the Future of "Integrated Medicine," https: / / www.mhlw.go.jp / stf / shingi / other-isei_127369.html Reference website 2: Japan Society of Oriental Medicine Homepage / Chinese medicine examination, http: / / www.jsom.or.jp / universally / examination / index.html

[0019] Currently, one of the challenges in promoting integrated medical care combining Western medicine and traditional Chinese medicine is the current state of traditional Chinese medicine prescriptions, which are based on the qualitative diagnostic methods of traditional Chinese medicine and statistics that are difficult to standardize.

[0020] Unlike Western medicine, which is based on scientific classification of diseases and pathological elucidation, the qualitative diagnostic method of Kampo medicine is based on subjective, relative diagnosis, which captures imbalances from each patient's normal state. Therefore, a more appropriate approach is to use statistical methods to accumulate data and mathematically model the diagnostic results of Kampo practitioners, who have a high level of empirical knowledge. However, diagnostic biases inherent in Kampo practitioners' diagnoses, and the statistical methods themselves are currently poorly understood. For this reason, expectations are high for automated diagnosis based on a database constructed using accurate waveform information indicating the pulse wave, noise removal from the acquired waveform information, feature extraction from multiple consecutive waveforms, and related information (e.g., diagnostic results from interviews, facial examinations, tongue examinations, abdominal examinations, etc.).

[0021] On the other hand, the difficulty in standardizing statistics lies in the fact that the nature of the mild, long-term improvement effects of herbal medicines makes them difficult to distinguish from placebo effects. It is thought that an approach involving the accumulation of statistical data on diagnosis / prescription based on the experience of skilled herbalists and modeling would be appropriate, but this is still not fully understood.

[0022] The following three references, Reference 1 to Reference 3, are cited as references for pulse wave diagnosis.

[0023] Reference 1: Japanese Patent Application Laid-Open No. 2009-011585 Reference 2: Japanese Patent Application Laid-Open No. 2004-195204 Reference 3: JP 2020-108819 A

[0024] Reference 1 describes a wristwatch-type 24-hour wearable pulse wave monitoring device that performs health management based on exercise and heart rate.

[0025] Reference 2 describes a device that detects indicators that represent the characteristics of blood pressure waveforms and determines the prescription of Western medicines such as calcium channel blockers and beta blockers based on the systolic blood pressure and AI (Augmentation Index) values.

[0026] Reference 3 describes a device that estimates blood glucose levels by utilizing the correlation between the AI ​​value of a blood pressure waveform and postprandial blood glucose levels.

[0027] As described above, the devices described in References 1 to 3 automatically perform physical condition monitoring, prescribing Western medicines, estimating blood glucose levels, etc. through automatic diagnostic analysis of pulse waves, but they are not yet capable of automatically prescribing herbal medicines based on automatic analysis of pulse waves.

[0028] Meanwhile, prior art for prescribing herbal medicines based on pulse wave diagnostic results includes efforts to have natural language processing AI (Artificial Intelligence) learn disease diagnoses and prescribed medications from electronic medical records that contain prescriptions for herbal medicines, and a turntable for prescribing herbal medicines that provides a quick reference table of diagnosis and prescribed medications based on the Chinese herbal medicine dialectical treatment method.

[0029] In recent years, information processing devices that use estimation models based on statistical learning have emerged to automate the prescription of herbal medicines. Details of this information processing device are described in Patent Document 1, cited as a prior art document. That is, as mentioned above, this information processing device outputs information indicating a prescription for a herbal medicine associated with the symptoms of a target subject whose symptoms are to be diagnosed by a herbal medicine doctor, based on a database containing information that associates, for each of a plurality of subjects, the symptoms of the subject with the history of the herbal medicine prescriptions given to the subject.

[0030] Here, the symptoms of the subject described in Patent Document 1 include headache, dizziness, menopausal disorder, etc., and refer to the physical or mental state of the subject diagnosed by a traditional Chinese medicine doctor. For this reason, with an information processing device such as that described in Patent Document 1, the traditional Chinese medicine doctor must diagnose the symptoms based on diagnostic methods such as interviewing, facial examination, tongue examination, abdominal examination, and pulse examination, and input information indicating the diagnosis results into the information processing device. For this reason, there are cases where the information processing device cannot sufficiently reduce the effort required for the traditional Chinese medicine doctor to prescribe traditional Chinese medicine to the subject.

[0031] Therefore, the information processing device of the embodiment is equipped with a prescription candidate output unit that outputs output information including herbal medicine candidate information indicating candidates for herbal medicine to be prescribed to the first subject, based on a first time-frequency spectrum image showing a first time-frequency spectrum corresponding to the waveform of the pulse wave of the first subject.

[0032] This allows the information processing device according to the embodiment to automate the process from pulse diagnosis to prescribing a herbal medicine, thereby reducing the time and effort required for a herbal medicine doctor to prescribe a herbal medicine to the first subject.

[0033] The configuration of the information processing device according to the embodiment and the processing performed by the information processing device will be described in detail below.

[0034] <Configuration of information processing device> The configuration of the information processing device according to the embodiment will be described below using the information processing device 20 as an example of the information processing device according to the embodiment. Note that in the embodiment, the symptoms of a subject refer to the physical or mental state of the subject diagnosed by a traditional Chinese medicine doctor. Therefore, in the embodiment, the symptoms of a certain subject do not include the subject's pulse wave and the waveform of that pulse wave.

[0035] FIG. 1 is a diagram illustrating an example of the configuration of an information processing system 1 including an information processing device 20. Here, the three-dimensional coordinate system TC is a three-dimensional Cartesian coordinate system that indicates directions in a diagram in which the three-dimensional coordinate system TC is drawn. In the following, for convenience of explanation, the X-axis in the three-dimensional coordinate system TC will be simply referred to as the X-axis. In the following, for convenience of explanation, the Y-axis in the three-dimensional coordinate system TC will be simply referred to as the Y-axis. In the following, for convenience of explanation, the Z-axis in the three-dimensional coordinate system TC will be simply referred to as the Z-axis. In the following, for convenience of explanation, the positive direction of the Z-axis will be referred to as the up or upward direction, and the negative direction of the Z-axis will be referred to as the down or downward direction.

[0036] The information processing system 1 includes a pulse wave detecting device 10 and an information processing device 20 which is an example of an information processing device according to an embodiment.

[0037] The pulse wave detecting device 10 detects the pulse wave of a subject. In this embodiment, the subject may be any person whose pulse wave is detected by the pulse wave detecting device 10.

[0038] Pulse wave detecting device 10 may have any configuration as long as it is capable of detecting the pulse wave of a subject. In the example shown in Fig. 1, pulse wave detecting device 10 includes a first member 11 on which one arm of the subject can be placed and fixed, a pulse wave sensor 12 that is brought into contact with the arm of the subject fixed by first member 11 to detect the pulse wave of the subject, and a second member 13 that supports pulse wave sensor 12.

[0039] The first member 11 is, for example, a platform on which one arm of the subject can be placed and fixed. The first member 11 may be configured to allow the position of the subject's arm relative to the pulse wave sensor 12 to be moved along a horizontal plane, or may be configured not to allow the position to be moved along a horizontal plane.

[0040] The pulse wave sensor 12 may be any sensor capable of detecting the pulse wave of the subject, such as a MEMS (Micro Electro Mechanical Systems) pressure sensor capable of detecting the pulse wave as a pressure fluctuation. The pulse wave sensor 12 is communicatively connected to the information processing device 20 via a wired or wireless connection. Thus, the pulse wave sensor 12 detects the pressure of the pulse wave and outputs an electrical signal corresponding to the detected pressure to the information processing device 20. As a result, the information processing device 20 can generate waveform information indicating the waveform of the subject's pulse wave during a predetermined measurement period based on the electrical signal acquired from the pulse wave sensor 12 during the measurement period.

[0041] Second member 13 may have any configuration as long as it is capable of supporting pulse wave sensor 12. Second member 13 may be configured to allow adjustment of the vertical position of pulse wave sensor 12 (i.e., the height of pulse wave sensor 12), or may be configured to prevent adjustment of the position.

[0042] In response to an operation received from the user, the information processing device 20 acquires an electrical signal from the pulse wave sensor 12 during a measurement period designated by the user. The information processing device 20 generates waveform information indicating the waveform of the subject's pulse wave based on the electrical signal acquired during the measurement period from the pulse wave sensor 12. The information processing device 20 stores the generated waveform information.

[0043] Based on the stored waveform information, the information processing device 20 calculates a time-frequency spectrum corresponding to the waveform indicated by the waveform information. Based on a time-frequency spectrum image indicating the calculated time-frequency spectrum, the information processing device 20 identifies candidate herbal medicines to be prescribed to the subject. After identifying the candidate herbal medicines, the information processing device 20 generates output information including candidate herbal medicine information indicating the identified candidate herbal medicines. After generating the output information, the information processing device 20 outputs the generated output information. This allows the information processing device 20 to reduce the effort required for a herbal medicine doctor to prescribe a herbal medicine to the subject.

[0044] Here, the information processing device 20 identifies candidate herbal medicines to be prescribed for the subject based on, for example, a first machine learning model that has been trained with first correspondence information in advance, a second machine learning model that has been trained with second correspondence information in advance, and the generated time-frequency spectral image. The first correspondence information is information in which a time-frequency spectral image corresponding to the waveform of the subject's pulse wave is associated with diagnostic result information indicating a diagnosis result of the subject by a traditional Chinese medicine doctor. The first machine learning model is a machine learning model that has been trained with the first correspondence information. When a time-frequency spectral image corresponding to the waveform of a subject's pulse wave is input, the first machine learning model outputs diagnostic result information indicating a diagnosis result that is estimated to be a plausible diagnosis result of the subject by the traditional Chinese medicine doctor. Meanwhile, the second correspondence information is information in which, for each subject, diagnostic result information indicating a diagnosis result of the subject by the traditional Chinese medicine doctor is associated with traditional Chinese medicine information indicating one or more traditional Chinese medicines prescribed for the subject by the traditional Chinese medicine doctor. The second machine learning model is a machine learning model that has been trained with the second correspondence information. When the second machine learning model receives the diagnostic result information output by the first machine learning model, it outputs herbal medicine candidate information indicating one or more herbal medicine candidates estimated to be plausible as one or more herbal medicines to be prescribed to the subject corresponding to the diagnostic result information. In this way, the information processing device 20 uses the first machine learning model and the second machine learning model to identify herbal medicine candidates to be prescribed to the subject. This allows the information processing device 20 to output herbal medicine candidate information indicating one or more herbal medicine candidates to be prescribed to the subject, while eliminating at least some of the subjective influence of the herbal medicine doctor. As a result, the information processing device 20 can reduce the effort required for the herbal medicine doctor to prescribe a herbal medicine to the subject, and can accurately prescribe the herbal medicine to the subject without being influenced by the herbal medicine doctor's experience. Details of the time-frequency spectral image and the diagnostic result information will be described later.

[0045] The information processing device 20 is, for example, a notebook PC (Personal Computer), a desktop PC, a workstation, a tablet PC, a multi-function mobile phone terminal (smartphone), a mobile phone terminal, a PDA (Personal Digital Assistant), etc., but is not limited to these.

[0046] <Hardware configuration of information processing device> The hardware configuration of the information processing device 20 will be described below with reference to Fig. 2. Fig. 2 is a diagram showing an example of the hardware configuration of the information processing device 20.

[0047] The information processing device 20 includes, for example, a processor 21, a storage unit 22, an input receiving unit 23, a communication unit 24, and a display unit 25. The information processing device 20 also communicates with the pulse wave detection device 10 via the communication unit 24. These components are connected to each other via a bus so that they can communicate with each other.

[0048] The processor 21 is, for example, a CPU (Central Processing Unit). Instead of a CPU, the processor 21 may be another processor such as an FPGA (Field Programmable Gate Array). The processor 21 executes various programs stored in the storage unit 22. The processor 21 may be configured by a CPU included in one information processing device (in this example, the information processing device 20), or may be configured by CPUs included in multiple information processing devices.

[0049] The storage unit 22 includes, for example, a hard disk drive (HDD), a solid state drive (SSD), an electrically erasable programmable read-only memory (EEPROM), a read-only memory (ROM), and a random access memory (RAM). The storage unit 22 may be an external storage device connected via a digital input / output port such as a universal serial bus (USB) instead of being built into the information processing device 20. The storage unit 22 stores various information, various programs, and the like processed by the information processing device 20. For example, the storage unit 22 stores the waveform information, the first machine learning model, the second machine learning model, and the like. The storage unit 22 may be configured with one storage device or multiple storage devices. The multiple storage devices may include a storage device provided in an information processing device separate from the information processing device 20.

[0050] The input receiving unit 23 is an input device such as a keyboard, a mouse, a touchpad, etc. The input receiving unit 23 may be a touch panel that is integrated with the display unit 25.

[0051] The communication unit 24 includes, for example, a digital input / output port such as a USB, an Ethernet (registered trademark) port, an antenna for communication, and the like.

[0052] The display unit 25 is, for example, a display panel such as a liquid crystal display panel or an organic EL (ElectroLuminescence) display panel.

[0053] <Functional configuration of information processing device> The functional configuration of the information processing device 20 will be described below with reference to Fig. 3. Fig. 3 is a diagram showing an example of the functional configuration of the information processing device 20.

[0054] The information processing device 20 includes a storage unit 22, an input receiving unit 23, a communication unit 24, a display unit 25, and a control unit 26.

[0055] The control unit 26 controls the entire information processing device 20. The control unit 26 includes an acquisition unit 261, a reception unit 262, a calculation unit 263, a prescription candidate output unit 264, a first learning unit 265, a second learning unit 266, a display control unit 267, and a generation unit 268. These functional units included in the control unit 26 are realized, for example, by the processor 21 executing various programs stored in the storage unit 22. Furthermore, some or all of the functional units may be hardware functional units such as LSI (Large Scale Integration) and ASIC (Application Specific Integrated Circuit).

[0056] The acquisition unit 261 acquires the electrical signal output from the pulse wave sensor 12 .

[0057] The reception unit 262 receives various types of information from the user.

[0058] The calculation unit 263 calculates various values ​​calculated by the information processing device 20. For example, based on certain waveform information, the calculation unit 263 calculates a time-frequency spectrum corresponding to the waveform indicated by the waveform information.

[0059] The prescription candidate output unit 264 generates the above-mentioned output information and outputs the generated output information. For example, the prescription candidate output unit 264 outputs the generated output information to the storage unit 22 and causes the storage unit 22 to store the output information.

[0060] The first learning unit 265 causes the first machine learning model to learn the first correspondence information.

[0061] The second learning unit 266 causes the second machine learning model to learn the second correspondence information.

[0062] The display control unit 267 generates various images and causes the display unit 25 to display the generated images.

[0063] The generating unit 268 generates waveform information indicating the waveform of the subject's pulse wave based on the electrical signal acquired by the acquiring unit 261. The generating unit 268 also generates a time-frequency spectrum image indicating the time-frequency spectrum calculated by the calculating unit 263.

[0064] <Processing by which information processing device generates waveform information> The process of generating waveform information by the information processing device 20 will be described below with reference to FIG. 4. FIG. 4 is a diagram showing an example of the process flow of generating waveform information by the information processing device 20. Hereinafter, a case will be described in which the information processing device 20 receives a first operation to cause the information processing device 20 to start the process of generating waveform information at a timing before the process of step S110 shown in FIG. 4 is performed. Hereinafter, a case will be described in which the information processing device 20 receives an operation to specify, as a measurement period, a period from the timing of receiving the first operation until a first time period specified by the user has elapsed. The first time period is, for example, one minute, but may alternatively be shorter or longer than one minute. Hereinafter, a case will be described in which the pulse wave detection device 10 starts detecting the pulse wave of the subject S1 at the timing. Here, when the pulse wave detection device 10 starts detecting the pulse wave of the subject S1, the pulse wave sensor 12 starts outputting the aforementioned electrical signal to the information processing device 20. In the following, as an example, a case will be described in which the information processing device 20 receives subject identification information for identifying the subject S1 at this timing.

[0065] After receiving the first operation, the acquisition unit 261 acquires electrical signals from the pulse wave sensor 12 at a predetermined sampling period during the measurement period until the first time has elapsed (step S110). In the embodiment, the process of converting the electrical signals from analog signals to digital signals is known, and therefore a description thereof will be omitted. Here, the sampling period is, for example, 0.002 seconds. Alternatively, the sampling period may be shorter or longer than 0.002 seconds. The acquisition unit 261 stores information indicating the electrical signals acquired during the measurement period in the storage unit 22.

[0066] Next, the generating unit 268 generates waveform information indicating the waveform of the pulse wave of the subject S1 within the measurement period based on the information indicating the electrical signal stored in the storage unit 22 in step S110 (step S120). Here, the method for generating the waveform information based on the information may be a known method or a method to be developed in the future. Furthermore, the information processing device 20 may be configured to perform the process of step S120 in parallel with the process of step S110.

[0067] Next, the generation unit 268 stores the waveform information generated in step S120 in the storage unit 22 (step S130). At this time, the generation unit 268 associates the waveform information with previously accepted subject identification information (i.e., subject identification information that identifies the subject S1) and stores the waveform information in the storage unit 22. After the process of step S130 is performed, the generation unit 268 ends the process of the flowchart shown in FIG. 4.

[0068] In this way, the information processing device 20 can store waveform information for each subject in the storage unit 22.

[0069] <Processing by an information processing device in which a first machine learning model learns first correspondence information and a second machine learning model learns second correspondence information> Hereinafter, with reference to FIG. 5, a process in which the information processing device 20 causes the first machine learning model to learn the first correspondence information and causes the second machine learning model to learn the second correspondence information will be described. FIG. 5 is a diagram showing an example of a process flow in which the information processing device 20 causes the first machine learning model to learn the first correspondence information and causes the second machine learning model to learn the second correspondence information. Hereinafter, as an example, a case will be described in which N pieces of waveform information generated by the processing of the flowchart shown in FIG. 4 are stored in the storage unit 22 at a timing before the processing of step S210 shown in FIG. 5 is performed. N may be any integer equal to or greater than 1. These N pieces of waveform information are waveform information for each of N subjects. Therefore, hereinafter, for convenience of explanation, each of these N subjects will be referred to as a training subject. Furthermore, hereinafter, for convenience of explanation, the waveform information for each of the N training subjects will be referred to as training waveform information. That is, hereinafter, as an example, a case will be described in which, at that timing, learning waveform information indicating the pulse wave waveforms of each of N learning subjects is stored in the storage unit 22. Also, hereinafter, as an example, a case will be described in which, at that timing, N pieces of diagnostic result information are stored in the storage unit 22. These N pieces of diagnostic result information are diagnostic result information for each of the N learning subjects. The diagnostic result information for a learning subject is information indicating the diagnosis result of the learning subject by a Chinese medicine doctor. In this case, the diagnostic result information for a learning subject is associated with subject identification information that identifies the learning subject. Also, hereinafter, as an example, a case will be described in which, at that timing, N pieces of herbal medicine information are stored in the storage unit 22. These N pieces of herbal medicine information are Chinese medicine information for each of the N learning subjects. The Chinese medicine information for a learning subject is information indicating one or more Chinese medicines prescribed by a Chinese medicine doctor for the learning subject. In this case, the herbal medicine information for a certain study subject is associated with subject identification information that identifies the study subject.

[0070] The control unit 26 reads out from the storage unit 22 each of the N pieces of learning-time waveform information stored in advance in the storage unit 22 (step S210). Note that in Fig. 5, the process of step S210 is indicated by "read waveform information".

[0071] Next, the control unit 26 selects, one by one, the subject identification information associated with each of the N pieces of learning-time waveform information read out in step S210 as target subject identification information, and repeats the processes of steps S230 to S270 for each selected piece of target subject identification information (step S220). Note that in Fig. 5, the process of step S220 is indicated by "for each subject identification information."

[0072] After the target subject identification information is selected in step S220, the first learning unit 265 refers to the N pieces of diagnostic result information stored in the memory unit 22 and reads out from the memory unit 22 the diagnostic result information associated with the target subject identification information selected in step S220 (step S230).

[0073] Here, the diagnostic result information will be explained. As mentioned above, the diagnostic result information for a certain study subject is information indicating the diagnostic result of the study subject by a Chinese medicine doctor. More specifically, the diagnostic result information is information indicating the results of the diagnosis of the study subject by the Chinese medicine doctor for each of M diagnostic items (for example, the results of pulse diagnosis, etc.). M may be any integer equal to or greater than 1. Therefore, each of the M diagnostic items includes multiple options that can be selected as a diagnostic result. Information indicating the diagnostic result for a certain diagnostic item among the M diagnostic items is represented by a vector having, as components, variables associated with each of the multiple options included in the diagnostic item. Therefore, the diagnostic result information is represented by the direct sum of vectors indicating the diagnostic results for each of the M diagnostic items. For example, if a diagnostic item among M diagnostic items includes three options X1 to X3, and if a Chinese medicine doctor selects X1 as the diagnostic result for that diagnostic item, the information indicating the diagnostic result is a vector having variables associated with X1 to X3 as components, with 1 assigned to the variable associated with X1 and 0 assigned to the variables associated with X2 and X3. Then, the diagnostic result information indicating the diagnostic results for each of the M diagnostic items is expressed by the direct sum of these six vectors.

[0074] In an embodiment, the M diagnostic items are the six major pulse types among the 28 disease pulses, namely, the floating network pulse, the sheng network pulse, the slow network pulse, the few network pulse, the virtual network pulse, and the substantial network pulse. In this case, M is 6. The multiple options included in the floating network pulse as options selectable by a Chinese medicine doctor as a diagnostic result are six pulse types: the floating network pulse, the kang network pulse (a Chinese character with a grass radical above a hole), the Hong network pulse, the leather network pulse, the wet network pulse, and the scattered network pulse. The multiple options included in the sheng network pulse as options selectable by a Chinese medicine doctor as a diagnostic result are four pulse types: the sheng network pulse, the fu network pulse, the weak network pulse, and the long network pulse. The multiple options included in the slow network pulse as options selectable by a Chinese medicine doctor as a diagnostic result are five pulse types: the slow network pulse, the brady network pulse, the slow network pulse, the jie network pulse, and the dai network pulse. The multiple options included in the few network pulse as options selectable by a Chinese medicine doctor as a diagnostic result are three pulse types: the rapid network pulse, the arterial network, and the rapid network pulse. The multiple options included in the Xuiwangmai as options selectable by a Chinese medicine doctor as a diagnostic result are four pulse types: Xuiwangmai, Short pulse, Small pulse, and Wei pulse. The multiple options included in the Shiwangmai as options selectable by a Chinese medicine doctor as a diagnostic result are six pulse types: Shimai, Long pulse, Xuanmai, Jinmai, Huamai, and Daimai. For example, information indicating the diagnostic result of a certain learning subject's Fubaimai is a vector having variables associated with each of the following: Fubaimai, Kangmai (a Chinese character with a grass crown above a hole), Hongmai, Kangmai, Weimai, and Sanmai as components. For example, if Fubaimai is selected by a Chinese medicine doctor as the diagnostic result for the Fubaimai, a component of the vector associated with the Fubaimai is assigned a value of 1. In this case, a value of 0 is assigned to each of the components of the vector associated with the Kangmai (a Chinese character with a grass crown above a hole), Hongmai, Kangmai, Weimai, and Sanmai. Here, certain diagnostic result information is represented by the direct sum of vectors indicating the diagnostic results for each of the six major vein types. That is, in the embodiment, the dimension of the vector representing certain diagnostic result information is 28. Note that, for convenience of explanation, the combination of vein types associated with each of the six components to which 1 is assigned among the components of the vector representing certain diagnostic result information will be referred to as a target vein type set. That is, the diagnostic result information is information indicating the target vein type set. Note that other information may be associated with this diagnostic result information.In the following, as an example, a case will be described in which the diagnostic result information is associated with information about the subject at the time of learning, information indicating the detection position of the pulse wave of the subject at the time of learning, and information indicating the medical history of the subject at the time of learning. Here, the information about the subject at the time of learning includes, for example, information indicating the gender, age, height, and weight of the subject at the time of learning.

[0075] Next, the first learning unit 265 refers to the N pieces of herbal medicine information stored in the storage unit 22, and reads out from the storage unit 22 the herbal medicine information associated with the target subject identification information selected in step S220 (step S240).

[0076] Next, the calculation unit 263 selects learning-time waveform information associated with the target subject identification information selected in step S220 from the learning-time waveform information read out in step S210, and calculates a time-frequency spectrum corresponding to the waveform indicated by the selected learning-time waveform information. Then, the generation unit 268 generates a time-frequency spectral image indicating the time-frequency spectrum calculated by the calculation unit 263 (step S250). For ease of explanation, the time-frequency spectral image indicating the time-frequency spectrum corresponding to the waveform indicated by the learning-time waveform information will be referred to as the learning-time time-frequency spectral image below.

[0077] Here, the time-frequency spectrum image will be described. FIG. 6 is a diagram showing an example of a waveform indicated by the learning waveform information associated with the target subject identification information selected in step S220. The vertical axis of the graph shown in FIG. 6 indicates the signal amplitude of the pulse wave. The horizontal axis of the graph indicates elapsed time. A curve plotted on the graph indicates the waveform. The calculation unit 263 removes frequency components equal to or higher than a predetermined frequency from the waveform indicated by the learning waveform information. The calculation unit 263 removes the frequency components using a band-pass filter. The information processing device 20 may be configured to receive a predetermined frequency from a user, or may be configured to receive a predetermined frequency by another method. The calculation unit 263 calculates a time-frequency spectrum based on information indicating the waveform after the frequency components have been removed. The calculation unit 263 calculates the time-frequency spectrum using an STFT (Short Time Fourier Transform). The generation unit 268 then generates a time-frequency spectrum image indicating the time-frequency spectrum calculated by the calculation unit 263. The time-frequency spectral image is, for example, a contour map in which spectral intensity is plotted as a function of time and frequency. FIG. 7 is a diagram showing an example of a time-frequency spectral image. The contour map shown in FIG. 7 is an example of a time-frequency spectral image for 5 seconds. The horizontal axis of the contour map represents time. The vertical axis of the contour map represents frequency. Each of the curves plotted on the contour map is a contour line of spectral intensity.

[0078] After the process of step S250 is performed, the first learning unit 265 generates first correspondence information (step S260). More specifically, the first learning unit 265 generates, as the first correspondence information, information that associates the learning time-frequency spectral image generated in step S250 with the diagnosis result information read out in step S230.

[0079] Next, the second learning unit 266 generates second correspondence information (step S270). More specifically, the second learning unit 266 generates information that associates the diagnosis result information read in step S230 with the herbal medicine information read in step S240 as the second correspondence information. The second learning unit 266 stores the generated second correspondence information in a second database pre-stored in the storage unit 22. That is, the second database is a database that stores N pieces of second correspondence information generated in the repeated processing of steps S220 to S270 shown in FIG. 5. The second database is a database with a two-dimensional table structure of n1 × m1. That is, the second database represents an n1 × m1 two-dimensional table. n1 is the number of combinations of vegetation types that can be selected as the target vegetation type set. n1 is 8640 because the number of vein types in the floating vein category is 6, the number of vein types in the sunken vein category is 4, the number of vein types in the slow vein category is 5, the number of vein types in the few vein category is 3, the number of vein types in the virtual vein category is 4, and the number of vein types in the real vein category is 6. m1 is the number of types of herbal medicines that the information processing device 20 can handle. The second learning unit 266 adds 1 to the value assigned to the field where the target vein type pair and each of the one or more herbal medicines indicated by the herbal medicine information intersect in the two-dimensional table, and stores the generated second correspondence information in the second database. Note that each field in the two-dimensional table is assigned a default value of 0. Figure 8 is a visual representation of an example of the processing in step S270.

[0080] After the process of step S270 is performed, the control unit 26 proceeds to step S220 and selects the next target subject identification information. Note that, if there is no unselected subject identification information in step S220, the control unit 26 ends the repeated process of steps S220 to S270.

[0081] After the repeated processing of steps S220 to S270 is completed, the first learning unit 265 causes the first machine learning model to learn the N pieces of first correspondence information stored in the first database in the repeated processing (step S280). Here, the processing of step S280 will be described.

[0082] In step S280, the first learning unit 265 trains a first machine learning model for each of the M diagnostic items. As a result, the first learning unit 265 can acquire a coefficient sequence of the trained first machine learning model for each of the M diagnostic items from the trained first machine learning model. In the embodiment, the M diagnostic items are each of the six major vein types, as described above. In this case, the first learning unit 265 trains a first machine learning model for each of the floating vein, the sinking vein, the slow vein, the few vein, the imaginary vein, and the real vein. For example, when training the first machine learning model for the floating vein, the first learning unit 265 extracts vectors indicating the diagnosis results of the floating vein and time-frequency spectral images associated with these vectors from each of the N pieces of first correspondence information. Then, for each extracted vector, the first learning unit 265 trains the first machine learning model with the time-frequency spectral image associated with the vector as the input and the vector as the output. When a time-frequency spectral image corresponding to the waveform of a subject's pulse wave is input to a first machine learning model trained with such inputs and outputs, as information indicating a diagnosis result for a floating pulse, a vector indicating the pulse type estimated to be most strongly present in the subject's pulse wave among pulse types included in the category of floating pulses. Here, the first machine learning model is a deep learning convolutional neural network (CNN). During training of the first machine learning model with such inputs and outputs, the input time-frequency spectral image is reduced in size using a kernel filter and single-output nonlinear processing, ultimately compressing it into one-dimensional data. During this training, the weights and biases of the first machine learning model are optimized so that the distribution of this one-dimensional data can be accurately distinguished to the number of options included in the diagnostic item (in this case, the number of pulse types included in the category of floating pulses, i.e., 6). By performing such training on the first machine learning model for each of the six major types, the first learning unit 265 can acquire a coefficient sequence of the trained first machine learning model from the trained first machine learning model for each of the M diagnostic items. Here, the coefficient sequence of a certain first machine learning model is a combination of the weights and biases of the first machine learning model.For ease of explanation, the coefficient sequence of the first machine learning model after learning about floating veins will be referred to as the first coefficient sequence. For ease of explanation, the coefficient sequence of the first machine learning model after learning about sinking veins will be referred to as the second coefficient sequence. For ease of explanation, the coefficient sequence of the first machine learning model after learning about slow veins will be referred to as the third coefficient sequence. For ease of explanation, the coefficient sequence of the first machine learning model after learning about few veins will be referred to as the fourth coefficient sequence. For ease of explanation, the coefficient sequence of the first machine learning model after learning about imaginary veins will be referred to as the fifth coefficient sequence. For ease of explanation, the coefficient sequence of the first machine learning model after learning about real veins will be referred to as the sixth coefficient sequence. When acquiring each of the first to sixth coefficient sequences, the first learning unit 265 may prepare a first machine learning model for each diagnostic item and train the first machine learning models in parallel for each diagnostic item, or may prepare a single first machine learning model and train the first machine learning model sequentially for each diagnostic item. After acquiring each of the first to sixth coefficient sequences in step S280, the first learning unit 265 stores coefficient sequence information indicating each of the first to sixth coefficient sequences in the storage unit 22. This allows the information processing device 20 to quickly reproduce the first machine learning model after training for floating net veins, for example, based on the first coefficient sequence and the first machine learning model.

[0083] Next, the second learning unit 266 trains the second machine learning model on the second database (i.e., the second correspondence information) generated in the repeated processing of steps S220 to S270 (step S290). More specifically, the second learning unit 266 trains the second machine learning model on the second database so that, when diagnosis result information indicating a target pulse type set for a certain subject is input, the second learning unit 266 outputs herbal medicine candidate information indicating one or more herbal medicines that are likely to be prescribed for the subject, as candidate herbal medicines to be prescribed for the subject. The machine learning model used as the second machine learning model may be any type of machine learning model as long as it can realize such input-output relationships. After the processing of step S290 is performed, the control unit 26 ends the processing of the flowchart shown in FIG. 5.

[0084] As described above, the information processing device 20 can cause the first machine learning model to learn the first correspondence information, and can cause the second machine learning model to learn the second correspondence information.

[0085] <Processing by which the information processing device receives diagnostic result information> Hereinafter, with reference to FIG. 9, a process in which the information processing device 20 receives diagnostic result information will be described. FIG. 9 is a diagram showing an example of the flow of a process in which the information processing device 20 receives diagnostic result information. Hereinafter, as an example, a case in which the information processing device 20 receives diagnostic result information about subject S2, who is one of the N learning subjects described above, will be described. Also, below, as an example, a case in which a diagnosis of subject S2 is made by a Chinese medicine doctor at a timing before the processing of step S310 shown in FIG. 9 is performed will be described. Also, below, as an example, a case in which the information processing device 20 receives a second operation that causes the information processing device 20 to start receiving diagnostic result information at that timing will be described.

[0086] After the information processing device 20 accepts the second operation, the display control unit 267 generates an information acceptance image PCT1 (step S310).

[0087] Here, the information reception image PCT1 is an image with which the information processing device 20 receives diagnostic result information. Fig. 10 is a diagram showing an example of the information reception image PCT1. The information reception image PCT1 includes, for example, eight images, namely, a first reception image G1 to an eighth reception image G8. Note that the information reception image PCT1 may be configured to include other images in addition to these eight images.

[0088] The first reception image G1 is a GUI that receives subject identification information. The first reception image G1 includes, for example, an input field into which the subject identification information is input.

[0089] The second reception image G2 is a GUI that receives information indicating the gender of the subject. The second reception image G2 includes, for example, two radio buttons: one for receiving information indicating that the gender of the subject is male, and the other for receiving information indicating that the gender of the subject is female.

[0090] The third reception image G3 is a GUI for receiving information indicating the age of the subject. The third reception image G3 includes, for example, an input field for inputting the information indicating the age of the subject.

[0091] The fourth reception image G4 is a GUI for receiving information indicating the height of the subject. The fourth reception image G4 includes, for example, an input field for inputting information indicating the height of the subject.

[0092] The fifth reception image G5 is a GUI for receiving information indicating the weight of the subject. The fifth reception image G5 includes, for example, an input field for inputting the information indicating the weight of the subject.

[0093] The sixth reception image G6 is a GUI that receives information indicating the position where the pulse wave of the subject is detected. The sixth reception image G6 includes, for example, a radio button for receiving information indicating that the arm where the pulse wave is detected is the left arm, a radio button for receiving information indicating that the arm where the pulse wave is detected is the right arm, a radio button for receiving information indicating that the position where the pulse wave of the subject is detected is 1 / 3, a radio button for receiving information indicating that the position where the pulse wave of the subject is detected is 2 / 3, and a radio button for receiving information indicating that the position where the pulse wave of the subject is detected is 3 / 4.

[0094] The seventh reception image G7 is a GUI for receiving information indicating the subject's medical history. The seventh reception image G7 includes, for example, an input field for inputting the information indicating the subject's medical history.

[0095] The eighth reception image G8 is a GUI for receiving diagnostic result information. The eighth reception image G8 includes, for example, six images, reception image G81 to reception image G86.

[0096] The reception image G81 is a GUI that accepts one of six vein types included in the category of floating net vein. When an operation to select the reception image G81 is performed, a drop-down menu L81 listing information indicating each of the six vein types is displayed. When the drop-down menu L81 is displayed, the user of the information processing device 20 can select one of the pieces of information indicating each of the six vein types listed in the drop-down menu L81. In the information reception image PCT1, when an operation to select one of the pieces of information indicating each of the six vein types listed in the drop-down menu L81 is performed, the display of the drop-down menu L81 disappears, and the information selected in the drop-down menu L81 is displayed in the display field of the reception image G81.

[0097] The reception image G82 is a GUI that accepts one of four vein types included in the category of sunken vein. When an operation to select the reception image G82 is performed, a drop-down menu L82 listing information indicating each of the four vein types is displayed. When the drop-down menu L82 is displayed, the user of the information processing device 20 can select one of the pieces of information indicating each of the four vein types listed in the drop-down menu L82. In the information reception image PCT1, when an operation to select one of the pieces of information indicating each of the four vein types listed in the drop-down menu L82 is performed, the display of the drop-down menu L82 disappears, and the information selected in the drop-down menu L82 is displayed in the display field of the reception image G82.

[0098] The reception image G83 is a GUI that accepts one of five vein types included in the category of slow-nettle vein. When an operation to select the reception image G83 is performed, a drop-down menu L83 listing information indicating each of the five vein types is displayed. When the drop-down menu L83 is displayed, the user of the information processing device 20 can select one of the pieces of information indicating each of the five vein types listed in the drop-down menu L83. In the information reception image PCT1, when an operation to select one of the pieces of information indicating each of the five vein types listed in the drop-down menu L83 is performed, the display of the drop-down menu L83 disappears, and the information selected in the drop-down menu L83 is displayed in the display field of the reception image G83.

[0099] The reception image G84 is a GUI that accepts one of three vein types included in the category of several vein types. When an operation to select the reception image G84 is performed, a drop-down menu L84 listing information indicating each of the three vein types is displayed. When the drop-down menu L84 is displayed, the user of the information processing device 20 can select one of the pieces of information indicating each of the three vein types listed in the drop-down menu L84. In the information reception image PCT1, when an operation to select one of the pieces of information indicating each of the three vein types listed in the drop-down menu L84 is performed, the display of the drop-down menu L84 disappears, and the information selected in the drop-down menu L84 is displayed in the display field of the reception image G84.

[0100] The reception image G85 is a GUI that accepts one of the four vein types included in the category of imaginary network veins. When an operation to select the reception image G85 is performed, a drop-down menu L85 listing information indicating each of the four vein types is displayed. When the drop-down menu L85 is displayed, the user of the information processing device 20 can select one of the pieces of information indicating each of the four vein types listed in the drop-down menu L85. In the information reception image PCT1, when an operation to select one of the pieces of information indicating each of the four vein types listed in the drop-down menu L85 is performed, the display of the drop-down menu L85 disappears, and the information selected in the drop-down menu L85 is displayed in the display field of the reception image G85.

[0101] The reception image G86 is a GUI that accepts one of six vein types included in the category of actual network veins. When an operation to select the reception image G86 is performed, a drop-down menu L86 listing information indicating each of the six vein types is displayed. When the drop-down menu L86 is displayed, the user of the information processing device 20 can select one of the pieces of information indicating each of the six vein types listed in the drop-down menu L86. In the information reception image PCT1, when an operation to select one of the pieces of information indicating each of the six vein types listed in the drop-down menu L86 is performed, the display of the drop-down menu L86 disappears, and the information selected in the drop-down menu L86 is displayed in the display field of the reception image G86.

[0102] 11 is a diagram showing an example of how each of the six drop-down menus described above is displayed. Here, the information reception image PCT1 may not include some or all of the second reception image G2 to the seventh reception image G7.

[0103] After the process of step S310 is performed, the display control unit 267 causes the display unit 25 to display the information acceptance image PCT1 generated in step S310 (step S320).

[0104] Next, the accepting unit 262 waits until the information processing device 20 accepts an operation via the information acceptance image PCT1 displayed on the display unit 25 in step S320 (step S330).

[0105] If the information processing device 20 receives an operation via the information reception image PCT1 displayed on the display unit 25 in step S320 (step S330-YES), the reception unit 262 determines whether an operation to terminate the reception of diagnostic result information via the information reception image PCT1 has been received in step S330 (step S340).

[0106] If the receiving unit 262 determines that an operation to terminate the reception of diagnostic result information via the information reception image PCT1 has not been received in step S330 (NO in step S340), the receiving unit 262 performs processing in accordance with the operation received in step S330 (step S370). Here, the processing is, for example, processing in which the information processing device 20 receives diagnostic result information via the eighth reception image G8. That is, through the processing in step S370, the information processing device 20 receives various information included in the diagnostic result information for the subject S2. Note that the processing in accordance with the operation received in step S330 may be any processing that can be performed in accordance with an operation received via the information reception image PCT1. After the processing in step S370 is performed, the receiving unit 262 transitions to step S330 and waits again until the information processing device 20 receives an operation via the information reception image PCT1 displayed on the display unit 25 in step S320.

[0107] On the other hand, if the receiving unit 262 determines that an operation to terminate the reception of diagnostic result information via the information reception image PCT1 has been received in step S330 (step S340-YES), the receiving unit 262 identifies a target vein type set based on the information received via each of the reception images G81 to G86 of the information reception image PCT1. Then, the receiving unit 262 generates diagnostic result information indicating the identified target vein type set (step S350). At this time, the receiving unit 262 associates the generated diagnostic result information with the information received via each of the first reception image G1 to seventh reception image G7 of the information reception image PCT1.

[0108] Next, the receiving unit 262 stores the diagnosis result information generated in step S350 in the storage unit 22 (step S360), and the process of the flowchart shown in FIG. 9 ends.

[0109] As described above, the information processing device 20 can receive the diagnosis result information.

[0110] <Processing by which an information processing device outputs herbal medicine candidate information> The process of the information processing device 20 outputting herbal medicine candidate information will be described below with reference to FIG. 12. FIG. 12 is a diagram showing an example of the process flow of the information processing device 20 outputting herbal medicine candidate information. The following describes, as an example, a case where first waveform information generated by the process of the flowchart shown in FIG. 4 is stored in the storage unit 22 at a timing before the process of step S410 shown in FIG. 12 is performed. The first waveform information is waveform information showing the waveform of the pulse wave of the subject S3. The following also describes, as an example, a case where, at that timing, the information processing device 20 receives subject identification information identifying the subject S3 as subject identification information identifying the subject to whom the herbal medicine candidate information is to be provided, along with a third operation that causes the information processing device 20 to start the process of outputting the herbal medicine candidate information. The following also describes, as an example, a case where, at that timing, coefficient sequence information is stored in the storage unit 22 by the process of the flowchart shown in FIG. 5.

[0111] After the information processing device 20 accepts the third operation and the subject identification information, the calculation unit 263 reads out the first waveform information from the storage unit 22 based on the accepted subject identification information (step S410). In Fig. 12, the process of step S410 is indicated by "read waveform information."

[0112] Next, the calculation unit 263 calculates a time-frequency spectrum according to the waveform indicated by the first waveform information read out in step S410. Then, the generation unit 268 generates a time-frequency spectrum image indicating the time-frequency spectrum calculated by the calculation unit 263 (step S420).

[0113] Next, the prescription candidate output unit 264 reads out, from the storage unit 22, coefficient sequence information that has been stored in advance in the storage unit 22. Then, the prescription candidate output unit 264 identifies candidates for herbal medicines to be prescribed to the subject S3 based on each of the first to sixth coefficient sequences indicated by the read coefficient sequence information and the time-frequency spectral image generated by the generation unit 268 in step S420 (step S430). In FIG. 12, the processing of step S430 is indicated by "identify candidate herbal medicine." Here, the processing of step S430 will be described.

[0114] The prescription candidate output unit 264 reproduces a trained first machine learning model for floating vein varices based on, for example, the first coefficient sequence among the first to sixth coefficient sequences indicated by the read coefficient sequence information and the first machine learning model. The prescription candidate output unit 264 inputs the time-frequency spectral image generated by the generation unit 268 in step S420 as input to the reproduced first machine learning model. The first machine learning model to which the time-frequency spectral image is input calculates likelihoods indicating the likelihood of floating vein varices as a diagnostic result for each of the six options included in floating vein varices, and outputs a vector indicating the option with the highest calculated likelihood as a vector indicating a diagnostic result for floating vein varices for the subject S3. Figure 13 is a diagram showing an example of the likelihoods for multiple options included in a diagnostic item such as floating vein varices. The prescription candidate output unit 264 acquires the vector output from the first machine learning model. The prescription candidate output unit 264 performs this process for each of the first to sixth coefficient sequences. As a result, the prescription candidate output unit 264 can acquire vectors indicating the diagnosis results of each of the six major types for the subject S3 from the first machine learning model. Then, the prescription candidate output unit 264 generates the direct sum of the six vectors output by the first machine learning model as the diagnostic result information for the subject S3. For ease of explanation, the vector indicating the diagnostic result information for the subject S3 will be referred to as vector Y below. For ease of explanation, each of the 28 components of vector Y will be represented by y1 to y28 below. In this case, the components contained in the vector indicating the diagnostic result of floating veins are y1 to y6, the components contained in the vector indicating the diagnostic result of sinking veins are y7 to y10, the components contained in the vector indicating the diagnostic result of slow veins are y11 to y15, the components contained in the vector indicating the diagnostic result of few veins are y16 to y18, the components contained in the vector indicating the diagnostic result of imaginary veins are y19 to y22, and the components contained in the vector indicating the diagnostic result of real veins are y23 to y28.

[0115] After the first machine learning model outputs the vector Y, the prescription candidate output unit 264 inputs the vector Y output from the first machine learning model as an input to the second machine learning model. As described above, when the second machine learning model receives the diagnostic result information output by the first machine learning model, it outputs herbal medicine candidate information indicating one or more herbal medicine candidates that are estimated to be likely to be prescribed to the subject S3 corresponding to the diagnostic result information. Therefore, when the second machine learning model receives the vector Y output by the first machine learning model, it outputs herbal medicine candidate information indicating one or more herbal medicine candidates that are estimated to be likely to be prescribed to the subject S3. Specifically, when the second machine learning model receives the vector Y, it identifies, based on the input vector Y and the previously trained second database (i.e., the second correspondence information), one or more fields associated with the target herbal medicine type pair indicated by the vector Y in the two-dimensional table indicated by the second database that are assigned a value equal to or greater than a predetermined first threshold. The second machine learning model then identifies the herbal medicines associated with each of the identified fields as one or more likely candidates for the herbal medicine to be prescribed to the subject S3. Here, the first threshold may be any value greater than 0. The first threshold is determined, for example, based on prior experimental results, so as to increase the accuracy of identifying one or more likely candidates for the herbal medicine to be prescribed to the subject S3. After identifying one or more likely candidates for the herbal medicine to be prescribed to the subject S3, the second machine learning model outputs herbal medicine candidate information indicating each of the identified one or more herbal medicine candidates. Here, FIG. 14 is a diagram visualizing the process by which the second machine learning model identifies one or more likely candidates for the herbal medicine to be prescribed to the subject S3.

[0116] After the second machine learning model outputs one or more pieces of herbal medicine candidate information, the prescription candidate output unit 264 identifies the herbal medicine candidates indicated by each of the one or more pieces of herbal medicine candidate information output by the second machine learning model as one or more herbal medicine candidates that are estimated to be plausible as the one or more herbal medicines to be prescribed to the subject S3. As described above, the prescription candidate output unit 264 identifies the one or more herbal medicine candidates in step S430.

[0117] After the processing of step S430 is performed, the prescription candidate output unit 264 generates output information including herbal medicine candidate information indicating each of the one or more herbal medicine candidates identified in step S430. The output information may include any information in addition to the herbal medicine candidate information indicating each of the one or more herbal medicine candidates. The prescription candidate output unit 264 outputs the generated output information to the display control unit 267. As a result, the display control unit 267 generates, for example, an image including the output information output from the prescription candidate output unit 264. The display control unit 267 then displays the generated image on the display unit 25 (step S440), and the processing of the flowchart shown in FIG. 12 ends. Note that in FIG. 12, the processing of step S440 is indicated by "display herbal medicine candidate information."

[0118] As described above, the information processing device 20 outputs output information including candidate herbal medicine information indicating candidate herbal medicines to be prescribed to the subject S3 based on the first waveform information indicating the waveform of the pulse wave of the subject S3. This allows the information processing device 20 to identify candidate herbal medicines to be prescribed to the subject S3 without having a herbal medicine doctor take a pulse diagnosis on the subject S3. As a result, the information processing device 20 can reduce the effort required for the herbal medicine doctor to prescribe the herbal medicine to the subject S3. This also allows the information processing device 20 to efficiently suppress variations in treatment results due to the proficiency of the herbal medicine doctor.

[0119] <First Modification of the Embodiment> A first modification of the embodiment will be described below. In this modification, the diagnostic result information for a certain subject includes information indicating at least one of the diagnostic results of the subject's interview, facial examination, tongue examination, and abdominal examination by a traditional Chinese medicine doctor. The following describes, as an example, a case in which the diagnostic result information for a certain subject includes information indicating the diagnostic result of the subject's interview by a traditional Chinese medicine doctor. In this case, the diagnostic result information for a certain subject includes, for example, information about the subject, information indicating the detection position of the subject's pulse wave, and information indicating the subject's medical history. That is, the information about the subject, the information indicating the detection position of the subject's pulse wave, and the information indicating the subject's medical history are each an example of information indicating the diagnostic result of the subject's interview by a traditional Chinese medicine doctor. In this case, the vector Y includes 10 components y29 to y38 in addition to the 28 components y1 to y28. Here, y29 is a variable that is assigned 1 if the subject is male and 0 if the subject is female. y30 is a variable that is assigned 1 if the subject is female and 0 if the subject is male. y31 is a variable that is assigned the subject's age. y32 is a variable that is assigned the subject's height. y33 is a variable that is assigned the subject's weight. y34 is a variable that is assigned 1 if the arm in which the pulse wave was detected is the left arm and 0 if the arm in which the pulse wave was detected is the right arm. y34 is a variable that is assigned 0 if the arm in which the pulse wave was detected is the right arm. y35 is a variable that is assigned 1 if the position in which the subject's pulse wave was detected is in inches and 0 if the position in which the subject's pulse wave was detected is in inches or feet. y36 is a variable to which 1 is assigned if the position at which the subject's pulse wave is detected is a kana, and 0 is assigned if the position at which the subject's pulse wave is detected is sun or shaku. y37 is a variable to which 1 is assigned if the position at which the subject's pulse wave is detected is shaku, and 0 is assigned if the position at which the subject's pulse wave is detected is sun or kana. y38 is a variable to which 1 is assigned if the subject has a pre-existing condition, and 0 is assigned if the subject does not have a pre-existing condition.The vector Y according to the first modification of the embodiment may include some of y29 to y38.

[0120] 10, the receiving unit 262 identifies a target pulse type set based on information received via each of the reception images G81 to G86 of the information reception image PCT1. Then, the receiving unit 262 generates, as diagnostic result information, information including diagnosis result information indicating the identified target pulse type set and information received via each of the second to seventh reception images G2 to G7. At this time, the receiving unit 262 associates the generated diagnostic result information with the subject identification information received via the first reception image G1 of the information reception image PCT1.

[0121] In this case, the second database is a database with a two-dimensional table structure of (n1 × n2) × m1. Here, n2 is the number of combinations of the 10 values ​​assigned to each of y29 to y38. This allows the information processing device 20 to output information including candidate herbal medicine information indicating candidate herbal medicines to be prescribed for a given subject based on a combination of information indicating the subject's target pulse type set, information about the subject, information indicating the detected position of the subject's pulse wave, and information indicating the subject's medical history. As a result, the information processing device 20 can increase the accuracy of identifying candidate herbal medicines to be prescribed for the subject, thereby more reliably reducing the effort required by a herbal medicine doctor to prescribe a herbal medicine to subject S3.

[0122] <Modification 2 of the embodiment> A second modification of the embodiment will be described below. In the second modification of the embodiment, the first machine learning model performs error correction processing in step S430 shown in FIG. 12. For example, as described above, the prescription candidate output unit 264 reproduces a trained first machine learning model for floating vein based on the first coefficient sequence among the first to sixth coefficient sequences indicated by the coefficient sequence information read in step S430 and the first machine learning model. Then, the prescription candidate output unit 264 inputs the time-frequency spectral image generated by the generation unit 268 in step S420 as input to the reproduced first machine learning model. The first machine learning model to which the time-frequency spectral image is input calculates likelihoods indicating the likelihood of each of the six options included in floating vein as a diagnostic result for floating vein. At this time, the first machine learning model performs error processing if all of the calculated six likelihoods are below a predetermined threshold. The error processing is, for example, a process of replacing these six likelihoods with the average value of these six likelihoods. When the first machine learning model performs error processing in this case, it outputs a vector in which 1 is assigned to all components as a vector indicating the diagnosis result of floating vein for subject S3. On the other hand, if at least one of the six calculated likelihoods is equal to or greater than a predetermined threshold, the first machine learning model outputs a vector indicating the option with the highest calculated likelihood as a vector indicating the diagnosis result of floating vein for subject S3. The prescription candidate output unit 264 acquires the vector output from the first machine learning model. The prescription candidate output unit 264 performs such processing for each of the first to sixth coefficient sequences. Here, for example, if at least one of the six vectors acquired from the first machine learning model is a vector in which 1 is assigned to all components, the prescription candidate output unit 264 terminates the processing of the flowchart shown in FIG. 12. Then, the display control unit 267 displays on the display unit 25 information indicating the occurrence of an error, as well as information prompting the user to re-acquire waveform information indicating the waveform of the pulse wave of subject S3. The above error processing may be replaced by other processing such as setting an error flag and not storing the error.Furthermore, the prescription candidate output unit 264 may continue the processing of the flowchart shown in FIG. 12 even if at least one of the six vectors acquired from the first machine learning model is a vector in which 1 is assigned to all components. In this case, n1 in the two-dimensional table indicated by the second database is replaced with the number of pulse type combinations selectable as a target pulse type set, rather than the number of pulse type combinations selectable as a combination of six or more pulse types from the 28 types of pathological pulses. In this case, the diagnosis result information is replaced with information indicating a combination of six or more pulse types, i.e., a vector indicating a combination of six or more pulse types. By performing these replacements, the information processing device 20 can continue the processing of the flowchart shown in FIG. 12 even in this case.

[0123] <Modification 3 of the embodiment> A third variation of the embodiment will be described below. In the third variation of the embodiment, the first machine learning model may be configured to output, as diagnostic result information, a vector Z indicating the likelihood calculated for each of the 28 disease pulses, together with the vector Y, in step S430 shown in FIG. 12 . In this case, the prescription candidate output unit 264 updates the values ​​of each field in the two-dimensional table indicated by the second correspondence information based on the combination of one or more herbal medicine candidates output from the second machine learning model by inputting the vector Y output from the first machine learning model to the second machine learning model in step S430, and the vector Z output from the first machine learning model. Specifically, the prescription candidate output unit 264 adds the likelihood indicated by the vector Z to the value of the field where each of the 28 disease pulses intersects with the one or more herbal medicine candidates in the two-dimensional table. For example, the prescription candidate output unit 264 adds the likelihood of the floating pulse, among the likelihoods indicated by the vector Z, to the value of the field where the floating pulse intersects with the one or more herbal medicine candidates. This allows the information processing device 20 to prevent a decrease in the accuracy of identifying one or more candidate herbal medicines that are estimated to be likely to be prescribed to the subject S3, due to variations in the diagnostic accuracy of the herbal medicine doctors. In this case, the first machine learning model may be configured to estimate at least one of the factor ranking and contribution of the herbal medicine prescription determining factors for each of the six major herbal medicine types obtained by analyzing the second correspondence information. In this case, the first machine learning model may, for example, multiply the likelihood calculated for each of the six major herbal medicine types by a weight based on at least one of the estimated factor ranking and contribution. Here, the weights multiplied by the likelihood calculated for each of the six major herbal medicine types are normalized so that the sum of all weights equals 1. For example, the first machine learning model determines a weight to be multiplied by the likelihood of each of the six options included in the Umimyaku based on at least one of the factor ranking and contribution estimated for the Umimyaku. The weights increase as the factor ranking and contribution increase. After determining the weights, the first machine learning model multiplies the determined weights by the likelihood of each of the six options included in the floating net vein.This allows the information processing device 20 to more reliably prevent a decrease in the accuracy of identifying one or more candidate herbal medicines that are estimated to be plausible as the one or more herbal medicines to be prescribed to the subject S3, due to variations in the diagnostic accuracy of pulse types among herbalists.

[0124] <Fourth Modification of the Embodiment> A fourth variation of the embodiment will be described below. In this fourth variation, instead of the 28 pathological pulses, the M diagnostic items are each of the six diagnostic items in Japanese pulse diagnosis. Again, M is 6. These six diagnostic items are the strength or weakness of a floating pulse versus a sinking pulse, the strength or weakness of a rapid pulse versus a slow pulse, the strength or weakness of a large pulse versus a small pulse, the strength or weakness of a weak pulse versus a strong pulse, the strength or weakness of a tight pulse versus a slow pulse, and the strength or weakness of a smooth pulse versus a slow pulse. In Japanese pulse diagnosis, floating pulses and sinking pulses are in an opposing relationship. Therefore, a Chinese medicine practitioner using Japanese pulse diagnosis diagnoses which of the floating and sinking pulses is more prominent in the subject's pulse wave. This is the diagnosis of the strength or weakness of a floating or sinking pulse. Similarly, in Japanese pulse diagnosis, the strength or weakness of a rapid pulse versus a slow pulse, the strong pulse versus a small pulse, the weak pulse versus a strong pulse, the tight pulse versus a slow pulse, and the smooth pulse versus a slow pulse are in opposing relationships. For this reason, in Japanese pulse diagnosis, the strength and weakness of a floating pulse and a sinking pulse, the strength and weakness of a rapid pulse and a slow pulse, the strength and weakness of a large pulse and a small pulse, the strength and weakness of a weak pulse and a strong pulse, the strength and weakness of a tight pulse and a slow pulse, and the strength and weakness of a smooth pulse and a slow pulse are each diagnosed as six diagnostic items. Below, as an example, a case will be described in which each of these six diagnostic items is diagnosed by a Chinese medicine doctor as one of five levels, from level 1 to level 5. In this case, multiple options included in a diagnostic item among these six diagnostic items are each level 1 to level 5. In this case, the strength and weakness of a floating pulse and a sinking pulse indicates a stronger floating pulse as the level value is lower, and a stronger sinking pulse as the level value is higher. In addition, in this case, the strength and weakness of a rapid pulse and a slow pulse indicates a stronger rapid pulse as the level value is lower, and a stronger slow pulse as the level value is higher. In addition, in this case, the strength and weakness of a large pulse and a small pulse indicates a stronger large pulse as the level value is lower, and a stronger small pulse as the level value is higher. In this case, the strength of the difference between the weak pulse and the active pulse is such that the lower the level value, the stronger the weak pulse, and the higher the level value, the stronger the active pulse.In this case, the strength of the difference between the tense pulse and the slow pulse is such that the lower the level value, the stronger the tense pulse, and the higher the level value, the stronger the slow pulse.In this case, the strength of the difference between the smooth pulse and the slow pulse is such that the lower the level value, the stronger the smooth pulse, and the higher the level value, the stronger the slow pulse.For example, information indicating the diagnosis result of the strength of the floating and sinking pulses of a certain study subject is a vector having variables corresponding to levels 1 to 5 as components. For example, if a Chinese medicine doctor selects level 1 as the diagnosis result of the strength of the floating and sinking pulses, a value of 1 is assigned to the variable corresponding to level 1 among the components of the vector. In this case, a value of 0 is assigned to the variables corresponding to levels 2 to 5 among the components of the vector. Here, the diagnostic result information according to the fourth modification of the embodiment is represented by the direct sum of vectors indicating the diagnostic results for each of these six diagnostic items. That is, in the fourth modification of the embodiment, the dimension of the vector representing certain diagnostic result information is 30. In the fourth modification of the embodiment, the diagnostic result information indicates a target level set instead of a target pulse type set.

[0125] The information processing device 20 accepts the diagnosis result information indicating such a target level set via an information acceptance image PCT2 instead of the information acceptance image PCT1 shown in Fig. 10. For this reason, the display control unit 267 generates the information acceptance image PCT2 in step S310 shown in Fig. 9. Here, Fig. 15 is a diagram showing an example of the information acceptance image PCT2.

[0126] The information reception image PCT2 includes, for example, the first to seventh reception images G1 to G7 and the ninth reception image G9 instead of the eighth reception image G8. Note that the information reception image PCT2 may include other images in addition to these images. Note that the first to seventh reception images G1 to G7 have already been described, so their description will be omitted.

[0127] The ninth reception image G9 is a GUI for receiving diagnostic result information. The ninth reception image G9 includes, for example, six images, reception image G91 to reception image G96.

[0128] The reception image G91 is a GUI that accepts the value of the strength level between the floating pulse and the sinking pulse. In the reception image G91, five radio buttons are arranged side by side between information indicating a floating pulse and information indicating a sinking pulse. These five radio buttons are, in order from the radio button closest to the information indicating a floating pulse to the information indicating a sinking pulse, a radio button associated with level 1, a radio button associated with level 2, a radio button associated with level 3, a radio button associated with level 4, and a radio button associated with level 5. When an operation to select one of these five radio buttons is performed, information indicating that the radio button has been selected is superimposed on the radio button selected by the operation. The information processing device 20 then accepts the value of the level associated with the radio button selected by the operation as the value of the strength level between the floating pulse and the sinking pulse. For example, when the information processing device 20 accepts an operation to select the radio button associated with level 1 in the reception image G91, it accepts level 1 as the value of the strength level between the floating pulse and the sinking pulse.

[0129] The reception image G92 is a GUI that accepts the value of the strength level between the rate pulse and the bradycardia. Five radio buttons are arranged in a row in the reception image G92 between information indicating a rate pulse and information indicating a bradycardia. These five radio buttons are, in order from the radio button closest to the information indicating a rate pulse to the information indicating a bradycardia, a radio button associated with level 1, a radio button associated with level 2, a radio button associated with level 3, a radio button associated with level 4, and a radio button associated with level 5. When an operation to select one of these five radio buttons is performed, information indicating that the radio button has been selected is superimposed on the radio button selected by the operation. The information processing device 20 then accepts the value of the level associated with the radio button selected by the operation as the value of the strength level between the rate pulse and the bradycardia. For example, when the information processing device 20 accepts an operation to select the radio button associated with level 1 in the reception image G92, it accepts level 1 as the value of the strength level between the rate pulse and the bradycardia.

[0130] The reception image G93 is a GUI that accepts the value of the strength level between the major pulse and the minor pulse. Five radio buttons are arranged in a row in the reception image G93 between information indicating a major pulse and information indicating a minor pulse. These five radio buttons are, in order from the radio button closest to the information indicating a major pulse to the information indicating a minor pulse, a radio button associated with level 1, a radio button associated with level 2, a radio button associated with level 3, a radio button associated with level 4, and a radio button associated with level 5. When an operation to select one of these five radio buttons is performed, information indicating that the radio button has been selected is superimposed on the radio button selected by the operation. The information processing device 20 then accepts the value of the level associated with the radio button selected by the operation as the value of the strength level between the major pulse and the minor pulse. For example, when the information processing device 20 accepts an operation to select the radio button associated with level 1 in the reception image G93, it accepts level 1 as the value of the strength level between the major pulse and the minor pulse.

[0131] The reception image G94 is a GUI that accepts the value of the strength level between the insufficient pulse and the active pulse. Five radio buttons are arranged in a row in the reception image G94 between information indicating an insufficient pulse and information indicating an active pulse. These five radio buttons are, in order from the radio button closest to the information indicating an insufficient pulse to the information indicating an active pulse, a radio button associated with level 1, a radio button associated with level 2, a radio button associated with level 3, a radio button associated with level 4, and a radio button associated with level 5. When an operation to select one of these five radio buttons is performed, information indicating that the radio button has been selected is superimposed on the radio button selected by the operation. The information processing device 20 then accepts the level value associated with the radio button selected by the operation as the value of the strength level between the insufficient pulse and the active pulse. For example, when the information processing device 20 accepts an operation to select the radio button associated with level 1 in the reception image G94, it accepts level 1 as the value of the strength level between the insufficient pulse and the active pulse.

[0132] The reception image G95 is a GUI that accepts the value of the strength level between a tense pulse and a brady pulse. Five radio buttons are arranged in a row in the reception image G95 between information indicating a tense pulse and information indicating a brady pulse. These five radio buttons are, in order from the radio button closest to the information indicating a tense pulse to the information indicating a brady pulse, a radio button associated with level 1, a radio button associated with level 2, a radio button associated with level 3, a radio button associated with level 4, and a radio button associated with level 5. When an operation to select one of these five radio buttons is performed, information indicating that the radio button has been selected is superimposed on the radio button selected by the operation. As a result, the information processing device 20 accepts the value of the level associated with the radio button selected by the operation as the value of the strength level between a tense pulse and a brady pulse. For example, when the information processing device 20 accepts an operation to select the radio button associated with level 1 in the reception image G95, it accepts level 1 as the value of the strength level between a tense pulse and a brady pulse.

[0133] The reception image G96 is a GUI that accepts the value of the strength level between Hua and Shibu pulses. In the reception image G96, five radio buttons are arranged in a row between information indicating a Hua pulse and information indicating a Shibu pulse. These five radio buttons are, in order from the radio button closest to the information indicating a Hua pulse to the information indicating a Shibu pulse, a radio button associated with level 1, a radio button associated with level 2, a radio button associated with level 3, a radio button associated with level 4, and a radio button associated with level 5. When an operation to select one of these five radio buttons is performed, information indicating that the radio button has been selected is superimposed on the radio button selected by the operation. The information processing device 20 then accepts the value of the level associated with the radio button selected by the operation as the value of the strength level between Hua and Shibu pulses. For example, when the information processing device 20 accepts an operation to select the radio button associated with level 1 in the reception image G96, it accepts level 1 as the value of the strength level between Hua and Shibu pulses.

[0134] The information processing device 20 can accept diagnostic result information according to the fourth modification of the embodiment via the information acceptance image PCT2 configured as described above. The information processing device 20 then performs the process of the flowchart shown in FIG. 5 using the diagnostic result information accepted via the information acceptance image PCT2 configured as described above, thereby generating first and second correspondence information, and training a first machine learning model based on the generated first correspondence information and a second machine learning model based on the generated second correspondence information, as in the embodiment. As a result, the information processing device 20 can perform the process of the flowchart shown in FIG. 12, as in the embodiment, thereby reducing the effort required for a Chinese medicine doctor to prescribe a Chinese medicine to a first subject. Note that n1 in the second database storing the second correspondence information according to the fourth modification of the embodiment is the number of combinations of level values ​​selectable as a target level set. And n1 is 15625 because there are 5 levels included in each of the strengths and weaknesses of a floating pulse and a sinking pulse, a rapid pulse and a slow pulse, a large pulse and a small pulse, a weak pulse and a strong pulse, a tense pulse and a slow pulse, and a smooth pulse and a slow pulse.

[0135] <Fifth Modification of the Embodiment> The following describes the fifth variation of the embodiment. The fifth variation of the embodiment is a variation of the fourth variation of the embodiment. In the fifth variation of the embodiment, the first learning unit 265 corrects each of the N pieces of diagnostic result information read out in step S230 shown in Fig. 5 using a weight vector to reduce the variation in diagnosis due to the subjective judgment of the Chinese medicine doctor.

[0136] This correction will be described in detail below. In the fifth modification of the embodiment, a correction filter for each Chinese medicine doctor is stored in the storage unit 22. In this case, each of the N pieces of diagnostic result information stored in the storage unit 22 is associated with Chinese medicine doctor identification information that identifies the Chinese medicine doctor who diagnosed the diagnosis indicated by the respective diagnostic result information.

[0137] A certain Chinese medicine doctor's correction filter has weight vectors associated with the strengths and weaknesses of the floating and shen pulses, the strengths and weaknesses of the rapid and slow pulses, the strengths and weaknesses of the large and small pulses, the strengths and weaknesses of the weak and strong pulses, the strengths and weaknesses of the tight and slow pulses, and the strengths and weaknesses of the smooth and slow pulses. For example, the weight vector associated with the strengths and weaknesses of the floating and shen pulses is a vector that corrects a vector indicating the diagnosis by the Chinese medicine doctor regarding the strengths and weaknesses of the floating and shen pulses. More specifically, the weight vector is a vector having five weights of different magnitudes as components that are multiplied by five variables contained as components in the vector indicating the diagnosis by the Chinese medicine doctor regarding the strengths and weaknesses of the floating and shen pulses. The vector indicating the diagnosis by the Chinese medicine doctor regarding the strengths and weaknesses of the floating and shen pulses is corrected by the Hadamard product of the vector indicating the diagnosis by the Chinese medicine doctor regarding the strengths and weaknesses of the floating and shen pulses and the weight vector associated with the strengths and weaknesses of the floating and shen pulses. Here, the Hadamard product is the product of each element of a matrix, and is sometimes called the Schur product. That is, the Hadamard product of the vector and the weight vector is the Hadamard product of two 5-row, 1-column matrices. By such correction, the value of the variable assigned with 1 is corrected in the vector indicating the diagnosis result by the Chinese medicine doctor regarding the strength of the floating pulse and the sinking pulse. This situation is similar for the strength of the rapid pulse and the slow pulse, the strength of the large pulse and the small pulse, the strength of the weak pulse and the strong pulse, the strength of the fast pulse and the slow pulse, and the strength of the smooth pulse and the slow pulse.

[0138] 5, the first learning unit 265 identifies a correction filter of the Chinese medicine doctor identified by the Chinese medicine doctor identification information, based on the Chinese medicine doctor identification information associated with the diagnosis result information read from the storage unit 22. The first learning unit 265 corrects the vector indicated by the diagnosis result information, based on the identified correction filter. The first learning unit 265 then generates, as first correspondence information, information associating the corrected vector with the learning time-frequency spectral image generated in step S250. The second learning unit 266 also generates second correspondence information associating the corrected vector with each of the one or more Chinese medicines indicated in the Chinese medicine information read in step S240, and stores the generated second correspondence information in the second database. At this time, the second learning unit 266 identifies the aforementioned target level pair based on the corrected vector, and adds the values ​​of non-zero variables among the variables of the vector to the values ​​of the fields where the identified target level pair and each of the one or more candidate herbal medicines intersect in the second database. This allows the information processing device 20 to reduce the variability in diagnostics based on the subjective judgment of the herbal medicine doctor for the second database. As a result, the information processing device 20 can more reliably reduce the effort required for the herbal medicine doctor to prescribe the herbal medicine to the first subject.

[0139] <Sixth Modification of the Embodiment> The following describes a sixth variation of the embodiment. The sixth variation of the embodiment is a variation of the fourth variation of the embodiment. In the sixth variation of the embodiment, the second machine learning model has a herbal medicine contraindication filter. As described above, when the second machine learning model receives the diagnostic result information output by the first machine learning model, it outputs herbal medicine candidate information indicating one or more herbal medicine candidates that are estimated to be likely to be prescribed to the subject corresponding to the diagnostic result information. The herbal medicine contraindication filter is a filter that prevents the one or more herbal medicine candidates indicated in the herbal medicine candidate information output from the second machine learning model from including any contraindicated herbal medicine combinations.

[0140] The following describes the process in which the second machine learning model uses the herbal medicine contraindication filter. For example, when a vector indicating diagnostic result information for a certain subject is input, the second machine learning model, based on the input vector and a previously trained second database (i.e., second correspondence information), identifies one or more fields assigned a value equal to or greater than a predetermined first threshold among the fields associated with the target level set indicated by the vector in the two-dimensional table indicated by the second database. The value of a certain field among these one or more fields is treated as a likelihood indicating the likelihood that the herbal medicine associated with the field is the herbal medicine to be prescribed for the subject. Therefore, the herbal medicine contraindication filter selects these one or more fields one by one as target fields in descending order of the assigned likelihood. Then, the herbal medicine contraindication filter performs the following process for each selected target field. The herbal medicine contraindication filter identifies, as contraindication fields, fields associated with one or more herbal medicines that are contraindicated in combination with the herbal medicine associated with the selected target field, and identifies, as other fields, one or more fields other than the target field and the contraindication field. The herbal medicine contraindication filter then multiplies the value of the target field by 1, the value of the contraindication field by -1, and the values ​​of the other fields by 0. The herbal medicine contraindication filter performs this process for each selected target field, and can ultimately cause the second machine learning model to identify the herbal medicine combinations associated with each of the one or more fields to which positive values ​​are assigned as combinations of one or more herbal medicine candidates that are estimated to be plausible as one or more herbal medicines to be prescribed for the subject, and that do not include any contraindicated herbal medicine combinations. In this way, the herbal medicine contraindication filter is a filter that excludes contraindicated herbal medicine combinations from among the one or more herbal medicine candidates indicated by the herbal medicine candidate information.

[0141] <Seventh Modification of the Embodiment> A seventh variation of the embodiment will be described below. In the seventh variation, the pulse wave waveform of the subject S3 is detected by the pulse wave sensor 12 under conditions in which the pressure with which the pulse wave sensor 12, which detects the pulse wave waveform of the subject S3, is pressed against the subject S3 is not constant within a predetermined measurement time. Even in this case, the information processing device 20 can accurately identify candidates for herbal medicines to be prescribed for the subject. This means that even if the way in which the pulse wave sensor 12 is pressed against the subject changes each time a pulse diagnosis is performed, an appropriate herbal medicine can be prescribed for the subject.

[0142] To achieve this, the storage unit 22 of the information processing device 20 stores waveform information indicating a waveform detected by the following pressure-varied arterial wave detection method, as shown in the flowchart of FIG. 4. In the pressure-varied arterial wave detection method, the waveform of the subject's pulse wave is detected by the pulse wave sensor 12 while changing the pressure with which the pulse wave sensor 12, which detects the waveform of the subject's pulse wave, is pressed against the subject during a predetermined measurement time. In this case, the pulse wave detection device 10 includes at least one of a first member 11 configured to be able to move one arm of the subject up and down using an actuator or the like, and a second member 13 configured to be able to move the pulse wave sensor 12 up and down using an actuator or the like. This allows the pulse wave detection device 10 to cause the pulse wave sensor 12 to detect the pressure of the pulse wave while, for example, intermittently or continuously increasing the pressure with which the pulse wave sensor 12 is pressed against the subject. The following describes, as an example, a case in which pulse wave detection device 10 causes pulse wave sensor 12 to detect the pressure of a pulse wave while continuously increasing the pressure with which pulse wave sensor 12 is pressed against the subject in a range of 40 gf to 300 gf. Note that pulse wave detection device 10 may also be configured to cause pulse wave sensor 12 to detect the pressure of a pulse wave while intermittently or continuously decreasing the pressure with which pulse wave sensor 12 is pressed against the subject. FIG. 16 shows an example of the process by which information processing device 20 generates a time-frequency spectral image based on waveform information indicating the waveform of a pulse wave detected by the pressure-varied arterial wave detection method. Waveform WP1 shown in FIG. 16 represents an example of the waveform of a pulse wave detected by the pressure-varied arterial wave detection method. In step S110 shown in FIG. 4, information processing device 20 acquires from pulse wave sensor 12 an electrical signal corresponding to the pressure of the pulse wave detected by the pressure-varied arterial wave detection method. Then, the information processing device 20 generates waveform information indicating the waveform of the subject's pulse wave during the measurement period based on the electrical signal thus acquired during the measurement period. Thereafter, in step S250 shown in Fig. 5, the information processing device 20 generates a time-frequency spectral image based on the waveform information thus generated.In this case, the information processing device 20 uses, for example, a band-pass filter to remove frequency components above a predetermined frequency from the waveform indicated by the waveform information, and calculates a time-frequency spectrum by STFT based on information indicating the waveform after removing the frequency components. As a result, the information processing device 20 can generate a time-frequency spectral image indicating the calculated time-frequency spectrum. Image WP2 shown in FIG. 16 is an example of a time-frequency spectral image generated in this manner.

[0143] The information processing device 20 generates first correspondence information, which is information associating the time-frequency spectral image thus generated with the diagnostic result information received by the processing of the flowchart shown in FIG. 9 , and trains the generated first correspondence information in a first machine learning model. Therefore, the first machine learning model is trained to output appropriate diagnostic result information according to the time-frequency spectral image according to the seventh modification of the embodiment. As a result, the information processing device 20 can accurately identify an appropriate herbal medicine candidate to be prescribed for the subject S3, even if the pulse wave waveform of the subject S3 is detected by the pulse wave sensor 12 under conditions in which the pressure applied to the subject S3 by the pulse wave sensor 12 is not constant within a predetermined measurement time. In other words, the information processing device 20 can eliminate ambiguity caused by the amount of pressure applied by a herbal medicine doctor during pulse diagnosis, and can accurately identify an appropriate herbal medicine candidate to be prescribed for the subject S3.

[0144] The above-described items may be combined in any manner. Furthermore, the method of classifying pulse types according to the state of the organs described above may be another classification method such as 38 pathological pulses instead of 28 pathological pulses. In this case, the target pulse type set may be a combination of seven or more pulse types. However, even in this case, the overall flow of the processing performed by the information processing device 20 remains the same as the flow described above.

[0145] <Additional Notes> [1] An information processing device (information processing device 20 in the example described above) that includes a prescription candidate output unit that outputs output information including herbal medicine candidate information indicating candidates for herbal medicines to be prescribed 9 to a first subject (subject S3 in the example described above) based on a first time-frequency spectrum image that shows a first time-frequency spectrum corresponding to the waveform of the pulse wave of the first subject.

[0146] [2] The information processing device described in [1], further comprising: a calculation unit (in the example described above, the calculation unit 263) that calculates the first time-frequency spectrum based on first waveform information that indicates the waveform of the pulse wave of the first subject as a first waveform; and a generation unit (in the example described above, the generation unit 268) that generates the first time-frequency spectrum image based on the first time-frequency spectrum.

[0147] [3] The information processing device described in [2], wherein the calculation unit removes frequency components equal to or higher than a predetermined frequency from the first waveform, generates information indicating the first waveform after the frequency components have been removed as the first waveform information, and calculates the first time-frequency spectrum based on the generated first waveform information.

[0148] [4] The information processing device described in [2] or [3], wherein the prescription candidate output unit identifies candidate herbal medicines to be prescribed to the first subject based on a first machine learning model, a second machine learning model, and the first time-frequency spectral image, wherein the first machine learning model is a machine learning model that has been trained with first correspondence information in which a second time-frequency spectral image showing a second time-frequency spectrum corresponding to the waveform of the pulse wave of the second subject (in the example described above, the learning subject) is associated with second diagnosis result information indicative of a diagnosis result of the second subject by a Chinese medicine doctor, and the second machine learning model is a machine learning model that has been trained with second correspondence information in which the second diagnosis result information is associated with Chinese medicine information, and the Chinese medicine information is information indicative of each of one or more Chinese medicines prescribed by the Chinese medicine doctor to the second subject.

[0149] [5] The information processing device described in [4], wherein the first machine learning model is a convolutional neural network for deep learning.

[0150] [6] The calculation unit calculates the second time-frequency spectrum based on second waveform information showing a waveform of the pulse wave of the second subject as a second waveform, the generation unit generates the second time-frequency spectrum image based on the second time-frequency spectrum, and the information processing device includes a reception unit (reception unit 262 in the example described above) that receives the second diagnosis result information and the one or more pieces of herbal medicine information, and associates the second time-frequency spectrum image generated by the generation unit with the second diagnosis result information received by the reception unit. The information processing device described in [4] further comprises: a first learning unit (in the example described above, the first learning unit 265) that generates information that associates the second diagnosis result information received by the reception unit with the one or more pieces of herbal medicine information received by the reception unit as the first correspondence information and has the first machine learning model learn the generated first correspondence information; and a second learning unit (in the example described above, the second learning unit 266) that generates information that associates the second diagnosis result information received by the reception unit with the one or more pieces of herbal medicine information received by the reception unit as the second correspondence information and has the second machine learning model learn the generated second correspondence information.

[0151] [7] The second diagnostic result information includes information indicating a diagnosis result (in the example described above, for example, floating pulse, level 1, etc.) of the second subject by a Chinese medicine doctor for a predetermined first diagnostic item (in the example described above, for example, floating pulse, strength between floating pulse and sink pulse, etc.) and information indicating a diagnosis result (in the example described above, for example, sink pulse, level 2, etc.) of the second subject by a Chinese medicine doctor for a predetermined second diagnostic item (in the example described above, for example, shenwang pulse, strength between rapid pulse and slow pulse, etc.), and the first learning unit trains the first machine learning model with the information indicating the diagnosis result of the second subject by the Chinese medicine doctor for the first diagnostic item and the second time-frequency spectral image based on the first correspondence information, obtains a coefficient sequence of the first machine learning model for the first diagnostic item as a first coefficient sequence, and learns the first time-frequency spectral image based on the first correspondence information. The information processing device of [6], further comprising: making the first machine learning model learn information indicating a diagnosis result of the second subject by a Chinese medicine doctor for a second diagnostic item and the second time-frequency spectral image; acquiring a coefficient sequence of the first machine learning model for the second diagnostic item as a second coefficient sequence; and making the first machine learning model output first diagnostic result information including information indicating a plausible diagnosis result as a diagnosis result of the first subject by a Chinese medicine doctor for the first diagnostic item and information indicating a plausible diagnosis result as a diagnosis result of the first subject by a Chinese medicine doctor for the second diagnostic item, based on the first coefficient sequence, the second coefficient sequence, the first machine learning model, and the first time-frequency spectral image; and identifying candidates for Chinese medicine to be prescribed to the first subject based on the first diagnostic result information output by the first machine learning model and the second machine learning model.

[0152] [8] The information processing device described in [7], wherein each of the first diagnostic item and the second diagnostic item is one of the 28 pathological pulses: floating net pulse, sunken net pulse, slow net pulse, few net pulse, imaginary net pulse, and solid net pulse, and are mutually different diagnostic items.

[0153] [9] The information processing device described in [7], wherein the first diagnostic item includes a plurality of first options selectable as a diagnostic result, and the second diagnostic item includes a plurality of second options selectable as a diagnostic result, and the first machine learning model calculates a likelihood indicating the likelihood of each of the plurality of first options as a diagnostic result by a Chinese medicine doctor based on the first coefficient sequence and the first time-frequency spectral image, and estimates information indicating a likely diagnostic result as a diagnostic result by a Chinese medicine doctor for the first subject for the first diagnostic item based on the calculated likelihood, and calculates a likelihood indicating the likelihood of each of the plurality of second options as a diagnostic result by a Chinese medicine doctor based on the second coefficient sequence and the first time-frequency spectral image, and estimates information indicating a likely diagnostic result as a diagnostic result by a Chinese medicine doctor for the first subject for the second diagnostic item based on the calculated likelihood.

[0154]

[10] The information processing device described in [7], wherein each of the first diagnostic item and the second diagnostic item is one of the strengths and weaknesses of a floating pulse and a sinking pulse in Japanese pulse diagnosis, a strength and weakness between a rapid pulse and a slow pulse, a strength and weakness between a large pulse and a small pulse, a strength and weakness between a weak pulse and a strong pulse, a strength and weakness between a fast pulse and a slow pulse, and a strength and weakness between a smooth pulse and a slow pulse, and are mutually different diagnostic items.

[0155]

[11] The information processing device described in [9], wherein the first diagnostic item includes a plurality of first options selectable as a diagnostic result, and the second diagnostic item includes a plurality of second options selectable as a diagnostic result, and the first machine learning model calculates a likelihood indicating the likelihood of each of the plurality of first options as a diagnostic result by a Chinese medicine doctor based on the first coefficient sequence, the first machine learning model, and the first time-frequency spectral image, and estimates information indicating a likely diagnostic result as a diagnostic result by a Chinese medicine doctor for the first subject for the first diagnostic item based on the calculated likelihood, and calculates a likelihood indicating the likelihood of each of the plurality of second options as a diagnostic result by a Chinese medicine doctor based on the second coefficient sequence, the first machine learning model, and the first time-frequency spectral image, and estimates information indicating a likely diagnostic result as a diagnostic result by a Chinese medicine doctor for the first subject for the second diagnostic item based on the calculated likelihood.

[0156]

[12] The information processing device described in

[10] , wherein the first machine learning model generates a first vector as information indicating a plausible diagnostic result as a diagnosis result of the first subject by the Chinese medicine doctor for the first diagnostic item, corrects the generated first vector by a Hadamard product with a weight vector corresponding to the first diagnostic item, generates a second vector as information indicating a plausible diagnostic result as a diagnosis result of the first subject by the Chinese medicine doctor for the second diagnostic item, corrects the generated second vector by a Hadamard product with a weight vector corresponding to the second diagnostic item, and outputs the first diagnostic result information including the corrected first vector and the corrected second vector.

[0157]

[13] The information processing device described in

[11] , wherein the first machine learning model performs error processing when the likelihood for each of the plurality of first options is all less than a predetermined threshold, or when the likelihood for each of the plurality of second options is all less than the threshold.

[0158]

[14] The information processing device described in

[11] , wherein the first machine learning model estimates at least one of the factor ranking and contribution of the herbal medicine prescription determining factors for each of the first diagnostic item and the second diagnostic item based on the second correspondence information, multiplies the likelihood for each of the multiple first options by a first weight based on the estimated at least one of the factors, and multiplies the likelihood for each of the multiple second options by a second weight based on the estimated at least one of the factors, and the prescription candidate output unit updates the second correspondence information based on the likelihood for each of the multiple first options, the first weight, the likelihood for each of the multiple second options, the second weight, and the herbal medicine candidate information output from the second machine learning model.

[0159]

[15] The information processing device described in

[10] , wherein the second machine learning model has a herbal medicine contraindication filter that excludes contraindicated combinations of herbal medicines from one or more herbal medicine candidates indicated by the herbal medicine candidate information.

[0160]

[16] An information processing device described in any one of [1] to

[15] , wherein the waveform of the pulse wave of the first subject is a waveform detected by a sensor that detects the waveform of the pulse wave of the first subject while changing the pressure with which the sensor is pressed against the first subject within a predetermined measurement time.

[0161]

[17] An information processing method comprising a prescription candidate output step of outputting output information including herbal medicine candidate information indicating candidates for herbal medicine to be prescribed to a first subject based on a first time-frequency spectrum image showing a first time-frequency spectrum corresponding to the waveform of the pulse wave of the first subject.

[0162]

[18] A program for causing a computer to execute a prescription candidate output step of outputting output information including herbal medicine candidate information indicating candidates for herbal medicine to be prescribed to a first subject based on a first time-frequency spectrum image showing a first time-frequency spectrum corresponding to the waveform of the pulse wave of the first subject.

[0163] The embodiments of this disclosure have been described in detail above with reference to the drawings, but the specific configuration is not limited to this embodiment, and may be changed, substituted, deleted, etc. as long as it does not deviate from the gist of this disclosure.

[0164] Furthermore, a program for implementing the functions of any of the components of the above-described device may be recorded on a computer-readable recording medium and loaded into a computer system for execution. Here, the device in question may be, for example, the pulse wave detection device 10 or the information processing device 20. The term "computer system" as used herein includes hardware such as an operating system (OS) and peripheral devices. The term "computer-readable recording medium" refers to portable media such as a flexible disk, a magneto-optical disk, a ROM, or a compact disk (CD-ROM), as well as storage devices such as a hard disk built into a computer system. The term "computer-readable recording medium" also includes devices that retain a program for a certain period of time, such as volatile memory within a computer system that acts as a server or client when the program is transmitted via a network such as the Internet or a communication line such as a telephone line.

[0165] The above program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by transmission waves in the transmission medium. Here, the "transmission medium" that transmits the program refers to a medium that has the function of transmitting information, such as a network such as the Internet or a communication line such as a telephone line. The program may also be a program for implementing some of the functions described above, or may be a so-called differential file or differential program that can implement the functions described above in combination with a program already recorded in the computer system. [Explanation of symbols]

[0166] 1… Pulse wave Information Processing Systems (Information Processing System) , 10... pulse wave detection device, 11... first member, 12... pulse wave sensor, 13... second member, 20... Pulse wave Information processing device (Information processing device) , 21...processor, 22...storage unit, 23...input reception unit, 24...communication unit, 25...display unit, 26...control unit, 261...acquisition unit, 262...reception unit, 263...calculation unit, 264...prescription candidate output unit, 265...first learning unit, 266...second learning unit, 267...display control unit, 268...generation unit, TC...three-dimensional coordinate system

Claims

1. A calculation unit that calculates a first time-frequency spectrum based on first waveform information that indicates a waveform of a pulse wave of a first subject as a first waveform; a generation unit that generates a first time-frequency spectrum image based on the first time-frequency spectrum; a prescription candidate output unit that outputs output information including herbal medicine candidate information indicating candidates of herbal medicines to be prescribed to the first subject based on the first time-frequency spectral image, the prescription candidate output unit identifies a candidate herbal medicine to be prescribed to the first subject based on a first machine learning model, a second machine learning model, and the first time-frequency spectral image; the first machine learning model is a machine learning model that has been trained with first correspondence information in which a second time-frequency spectrum image showing a second time-frequency spectrum corresponding to a waveform of a pulse wave of a second subject is associated with second diagnosis result information showing a diagnosis result of the second subject by a Chinese medicine doctor; The second machine learning model is a machine learning model that has learned second correspondence information in which the second diagnosis result information and the herbal medicine information are associated with each other, The herbal medicine information is information indicating each of one or more herbal medicines prescribed by the herbal medicine doctor to the second subject. Pulse wave information processing device.

2. the calculation unit removes frequency components equal to or higher than a predetermined frequency from the first waveform, generates information indicating the first waveform after the frequency components have been removed as the first waveform information, and calculates the first time-frequency spectrum based on the generated first waveform information. The pulse wave information processing device according to claim 1 .

3. The first machine learning model is a convolutional neural network for deep learning. The pulse wave information processing device according to claim 1 or 2.

4. the calculation unit calculates the second time-frequency spectrum based on second waveform information indicating a waveform of the pulse wave of the second subject as a second waveform; the generation unit generates the second time-frequency spectral image based on the second time-frequency spectrum; The information processing device includes: a receiving unit that receives the second diagnosis result information and the one or more pieces of herbal medicine information; a first learning unit that generates information that associates the second time-frequency spectral image generated by the generating unit with the second diagnostic result information received by the receiving unit as the first correspondence information, and causes the first machine learning model to learn the generated first correspondence information; a second learning unit that generates second correspondence information by associating the second diagnosis result information received by the receiving unit with the one or more pieces of herbal medicine information received by the receiving unit, and causes the second machine learning model to learn the generated second correspondence information; Further comprising: The pulse wave information processing device according to claim 1 or 2.

5. the second diagnostic result information includes information indicating a diagnostic result of the second subject by the Chinese medicine doctor for a predetermined first diagnostic item, and information indicating a diagnostic result of the second subject by the Chinese medicine doctor for a predetermined second diagnostic item, the first learning unit, based on the first correspondence information, causes the first machine learning model to learn information indicating a diagnosis result of the second subject by a Chinese medicine doctor for the first diagnostic item and the second time-frequency spectral image, and acquires a coefficient sequence of the first machine learning model for the first diagnostic item as a first coefficient sequence; based on the first correspondence information, causes the first machine learning model to learn information indicating a diagnosis result of the second subject by a Chinese medicine doctor for the second diagnostic item and the second time-frequency spectral image, and acquires a coefficient sequence of the first machine learning model for the second diagnostic item as a second coefficient sequence; the prescription candidate output unit causes the first machine learning model to output first diagnostic result information, the first diagnostic result including information indicating a plausible diagnostic result for the first subject by the Chinese medicine doctor for the first diagnostic item and information indicating a plausible diagnostic result for the first subject by the Chinese medicine doctor for the second diagnostic item, based on the first coefficient sequence, the second coefficient sequence, the first machine learning model, and the first time-frequency spectral image, and identifies a candidate Chinese herbal medicine to be prescribed to the first subject based on the first diagnostic result information output by the first machine learning model and the second machine learning model. The pulse wave information processing device according to claim 4 .

6. The first diagnostic item and the second diagnostic item are either a floating net pulse, a sunken net pulse, a slow net pulse, a few net pulses, an imaginary net pulse, or a real net pulse among 28 diseased pulses, and are mutually different diagnostic items. The pulse wave information processing device according to claim 5 .

7. the first diagnostic item includes a plurality of first options selectable as diagnostic results; the second diagnostic item includes a plurality of second options selectable as diagnostic results, the first machine learning model calculates a likelihood indicating the likelihood of each of the plurality of first options as a diagnosis result by a Chinese medicine doctor based on the first coefficient sequence and the first time-frequency spectral image, and estimates information indicating a likely diagnosis result as a diagnosis result by a Chinese medicine doctor for the first subject for the first diagnostic item based on the calculated likelihood; calculates a likelihood indicating the likelihood of each of the plurality of second options as a diagnosis result by a Chinese medicine doctor based on the second coefficient sequence and the first time-frequency spectral image, and estimates information indicating a likely diagnosis result as a diagnosis result by a Chinese medicine doctor for the first subject for the second diagnostic item based on the calculated likelihood; The pulse wave information processing device according to claim 5 .

8. The first diagnostic item and the second diagnostic item are each one of the strength and weakness of a floating pulse and a sinking pulse, the strength and weakness of a rapid pulse and a slow pulse, the strength and weakness of a large pulse and a small pulse, the strength and weakness of a weak pulse and a strong pulse, the strength and weakness of a tight pulse and a slow pulse, and the strength and weakness of a smooth pulse and a slow pulse in Japanese pulse diagnosis, and are mutually different diagnostic items. The pulse wave information processing device according to claim 5 .

9. the first diagnostic item includes a plurality of first options selectable as diagnostic results; the second diagnostic item includes a plurality of second options selectable as diagnostic results, the first machine learning model calculates a likelihood indicating the likelihood of each of the plurality of first options being a diagnosis result by a Chinese medicine doctor based on the first coefficient sequence, the first machine learning model, and the first time-frequency spectral image, and estimates information indicating a likely diagnosis result as a diagnosis result by a Chinese medicine doctor for the first subject for the first diagnostic item based on the calculated likelihood; and calculates a likelihood indicating the likelihood of each of the plurality of second options being a diagnosis result by a Chinese medicine doctor based on the second coefficient sequence, the first machine learning model, and the first time-frequency spectral image, and estimates information indicating a likely diagnosis result as a diagnosis result by a Chinese medicine doctor for the first subject for the second diagnostic item based on the calculated likelihood; The pulse wave information processing device according to claim 7 .

10. the first machine learning model generates a first vector as information indicating a plausible diagnosis result as a diagnosis result of the first subject by the Chinese medicine doctor for the first diagnostic item, corrects the generated first vector by a Hadamard product with a weight vector associated with the first diagnostic item, generates a second vector as information indicating a plausible diagnosis result as a diagnosis result of the first subject by the Chinese medicine doctor for the second diagnostic item, corrects the generated second vector by a Hadamard product with a weight vector associated with the second diagnostic item, and outputs the first diagnostic result information including the corrected first vector and the corrected second vector. The pulse wave information processing device according to claim 8 .

11. the first machine learning model performs error processing when all likelihoods for the plurality of first options are less than a predetermined threshold, or when all likelihoods for the plurality of second options are less than the threshold; The pulse wave information processing device according to claim 9 .

12. The first machine learning model estimates at least one of a factor ranking and a contribution degree of a Chinese herbal medicine prescription determining factor for each of the first diagnostic item and the second diagnostic item based on the second correspondence information, multiplies a likelihood for each of the plurality of first options by a first weight based on the estimated at least one, and multiplies a likelihood for each of the plurality of second options by a second weight based on the estimated at least one, The prescription candidate output unit updates the second correspondence information based on the likelihood for each of the plurality of first options, the first weight, the likelihood for each of the plurality of second options, the second weight, and the herbal medicine candidate information output from the second machine learning model. The pulse wave information processing device according to claim 9 .

13. The second machine learning model has a herbal medicine contraindication filter that excludes contraindicated combinations of herbal medicines from among the one or more herbal medicine candidates indicated by the herbal medicine candidate information. The pulse wave information processing device according to claim 8 .

14. The waveform of the pulse wave of the first subject is a waveform detected by a sensor that detects the waveform of the pulse wave of the first subject while changing the pressure with which the sensor is pressed against the first subject within a predetermined measurement time. The pulse wave information processing device according to claim 1 or 2.

15. In a computer, a calculation step of calculating a first time-frequency spectrum based on first waveform information indicating a waveform of the pulse wave of the first subject as a first waveform; generating a first time-frequency spectrum image based on the first time-frequency spectrum; a prescription candidate output step of outputting output information including herbal medicine candidate information indicating candidates of herbal medicines to be prescribed to the first subject based on the first time-frequency spectral image, the prescription candidate output step identifies a candidate herbal medicine to be prescribed to the first subject based on a first machine learning model, a second machine learning model, and the first time-frequency spectral image; the first machine learning model is a machine learning model that has been trained with first correspondence information in which a second time-frequency spectrum image showing a second time-frequency spectrum corresponding to a waveform of a pulse wave of a second subject is associated with second diagnosis result information showing a diagnosis result of the second subject by a Chinese medicine doctor; The second machine learning model is a machine learning model that has learned second correspondence information in which the second diagnosis result information and herbal medicine information are associated with each other, The herbal medicine information is information indicating each of one or more herbal medicines prescribed by the herbal medicine doctor to the second subject. Pulse wave information processing method.

16. On the computer, a calculation step of calculating a first time-frequency spectrum based on first waveform information indicating a waveform of the pulse wave of the first subject as a first waveform; generating a first time-frequency spectrum image based on the first time-frequency spectrum; a prescription candidate output step of outputting output information including herbal medicine candidate information indicating candidates of herbal medicines to be prescribed to the first subject based on the first time-frequency spectral image; A program for executing the prescription candidate output step identifies a candidate herbal medicine to be prescribed to the first subject based on a first machine learning model, a second machine learning model, and the first time-frequency spectral image; the first machine learning model is a machine learning model that has been trained with first correspondence information in which a second time-frequency spectrum image showing a second time-frequency spectrum corresponding to a waveform of a pulse wave of a second subject is associated with second diagnosis result information showing a diagnosis result of the second subject by a Chinese medicine doctor; The second machine learning model is a machine learning model that has learned second correspondence information in which the second diagnosis result information and herbal medicine information are associated with each other, The herbal medicine information is information indicating each of one or more herbal medicines prescribed by the herbal medicine doctor to the second subject. Pulse wave information processing program.

Images