Malfunction factor estimation device, malfunction factor estimation method, and program

Abstract

For example, even though a cause of depressive symptoms (first level) is mental fatigue (second level), the mental fatigue can be due to various causes, such as human relations (third level) or inadequate sleep (third level). The purpose of the present invention is to estimate a cause of a disorder even when the cause of a physical disorder has a multilevel structure, and the cause of the physical disorder is difficult to understand.　To achieve this purpose, in the present disclosure, a disorder cause estimation device for estimating the cause of a physical disorder in a user has: a shared cause-effect structure database as a graph database in which cause-effect relationships form a structure in which causes that result in a physical disorder in the user serve as nodes; and a user disorder cause searching unit that fits disorder determination results based on data of responses from a prescribed user to prescribed questions, or data of sensing relating to the prescribed user to the nodes that correspond to the prescribed questions to determine cause scores that relate to a disorder, and thereby estimates the cause of a physical disorder in the prescribed user.

Description

[Technical Field]

[0001] The present disclosure relates to a technology for searching for a main cause of a user's physical ailment among multiple underlying causes of the ailment. [Background technology]

[0002] It is important to understand the factors (main causes) of physical discomfort that are specific to each individual user, such as improving QOL (Quality of Life) and preventing lifestyle-related diseases, and to know the remedial actions that are appropriate for these factors.

[0003] Non-Patent Document 1 discloses that the relationship between a time when one is feeling unwell and the actions (causes) taken before and after the time is analyzed, and a recommended action to improve the physical condition is selected. [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] "Behavioral modification methods for improving QOL using artificial intelligence and life logs," 2017 (https: / / www.jstage.jst.go.jp / article / jahpp / 30 / 0 / 30_126 / _pdf / -char / ja) Summary of the Invention [Problem to be solved by the invention]

[0005] However, in Non-Patent Document 1, when the relationship between physical ailments and their causes is multi-stage, it is difficult to identify the cause (main cause) of the ailment. For example, even if the cause of depressive symptoms is mental fatigue, there are various possible causes of this mental fatigue, such as interpersonal relationships or lack of sleep.

[0006] The present invention has been made in consideration of the above points, and aims to estimate the cause of illness even when the cause of illness has a multi-stage structure and the cause of the physical illness is difficult to understand. [Means for solving the problem]

[0007] In order to solve the above problem, the invention of claim 1 is an illness factor estimation device that estimates the cause of a user's physical illness, and includes: a common causal structure database as a graph database in which causal relationships are structured, with each node representing a cause that results in the user's physical illness; and a user illness cause search unit that estimates the cause of the user's physical illness by applying data on answers from a predetermined user to each predetermined question or each illness judgment result based on sensing data related to the predetermined user to each node corresponding to the predetermined question, and finding each cause score related to the illness; the common causal structure database is a graph database in which a causal relationship is structured with the result of the user's physical illness, a first cause that causes the result, and a second cause that causes the first cause as the nodes, and each edge is a weight that indicates the relative relationship between the result of the illness and the first cause, and the relative relationship between the first cause and the second cause, and the user illness cause search unit calculates each cause score based on the determination results applied to each node and the weight associated with each node. This is a malfunction cause estimation device. [Effects of the Invention]

[0008] As described above, according to the present invention, even when the cause of the illness is multi-staged and the cause of the physical illness is difficult to understand, it is possible to estimate the cause of the illness. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a diagram illustrating the overall configuration of a communication system according to an embodiment of the present invention. [Figure 2] 1 is a diagram illustrating an electrical hardware configuration of a malfunction factor estimation device and a communication terminal according to an embodiment of the present invention. [Figure 3] FIG. 2 is a functional configuration diagram of the malfunction factor estimation device. [Figure 4] FIG. 1 is a conceptual diagram showing nodes and edges of a common causal structure DB as a graph DB. [Figure 5] This is a conceptual diagram of the properties of a common causal structure DB as a graph DB. [Figure 6] FIG. 10 is a conceptual diagram showing nodes and edges of a user information DB as a graph DB. [Figure 7] FIG. 10 is a conceptual diagram of an advice message DB. [Figure 8] FIG. 10 is a diagram illustrating a method for calculating a cause score. [Figure 9] 3 is a flowchart showing a process or operation executed by the malfunction factor estimating device. [Figure 10] 3 is a flowchart showing a process or operation executed by the malfunction factor estimating device. [Figure 11] FIG. 10 is a diagram illustrating an example of a display on a communication terminal. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0011] [System configuration of the embodiment] First, the overall configuration of the communication system of this embodiment will be explained with reference to Fig. 1. Fig. 1 is a diagram showing the overall configuration of the communication system according to this embodiment.

[0012] 1, a communication system 10 of this embodiment is constructed by a malfunction factor estimation device 30 and a communication terminal 50. The communication terminal 50 is managed and used by a user. The user refers to an advice message transmitted from the malfunction factor estimation device 30 to the communication terminal 50 and considers subsequent actions (such as improving lifestyle).

[0013] Furthermore, the malfunction factor estimating device 30 and the communication terminal 50 can communicate with each other via a communication network 100 such as the Internet. The communication network 100 may be connected wirelessly or by wire.

[0014] The malfunction factor estimation device 30 is configured by one or more computers. When the malfunction factor estimation device 30 is configured by multiple computers, it may be referred to as a "malfunction factor estimation device" or a "malfunction factor estimation system."

[0015] The illness factor estimation device 30 is a device that estimates the cause (main cause) of an individual user's physical illness, and even if the cause of the illness has a multi-stage structure, it estimates the cause of the illness based on the user's direct answers to predetermined questions or sensing data from a wearable computer or the like worn by the user. As described above, a multi-stage cause indicates a state in which, for example, mental fatigue (stage 2) is thought to be the cause of depressive symptoms (stage 1), and further, interpersonal relationships or lack of sleep (stage 3) are thought to be the causes of the mental fatigue (stage 2).

[0016] The communication terminal 50 is a computer, and a smartphone is shown as an example in Fig. 1. In Fig. 1, a user operates the communication terminal 50. Note that the malfunction factor estimation device 30 may perform processing independently without using the communication terminal 50.

[0017] [Hardware configuration] <Hardware configuration of malfunction factor estimation device> Next, the electrical hardware configuration of the malfunction factor estimating device 30 will be described with reference to Fig. 2. Fig. 2 is a diagram showing the electrical hardware configuration of the malfunction factor estimating device.

[0018] As shown in FIG. 2, the malfunction factor estimation device 30 is a computer that includes a CPU (Central Processing Unit) 301, a ROM (Read Only Memory) 302, a RAM (Random Access Memory) 303, an SSD (Solid State Drive) 304, an external device connection I / F (Interface) 305, a network I / F 306, a media I / F 309, and a bus line 310.

[0019] Of these, the CPU 301 controls the overall operation of the malfunction factor estimating device 30. The ROM 302 stores programs such as an IPL (Initial Program Loader) used to drive the CPU 301. The RAM 303 is used as a work area for the CPU 301.

[0020] The SSD 304 reads or writes various data under the control of the CPU 301. Note that instead of the SSD 304, a hard disk drive (HDD) may be used.

[0021] The external device connection I / F 305 is an interface for connecting various external devices, such as a display, a speaker, a keyboard, a mouse, a USB (Universal Serial Bus) memory, and a printer.

[0022] The network I / F 306 is an interface for performing data communication via the communication network 100 .

[0023] The media I / F 309 controls reading and writing (storing) of data from and to a recording medium 309m such as a flash memory, etc. The recording medium 309m includes a DVD (Digital Versatile Disc) and a Blu-ray Disc (registered trademark).

[0024] The bus line 310 is an address bus, a data bus, etc. for electrically connecting the components such as the CPU 301 shown in FIG.

[0025] The electrical hardware configuration of the communication terminal 50 is basically the same as that of the malfunction factor estimation device 30, and therefore a description thereof will be omitted.

[0026] [Functional configuration of the malfunction factor estimation device] Next, a description will be given of the functional configuration of the malfunction factor estimation device 30. Fig. 3 is a functional configuration diagram of the malfunction factor estimation device.

[0027] 3, the malfunction factor estimating device 30 has a user state acquiring unit 31, a user malfunction cause searching unit 33, an advice presenting unit 35, a common causal structure weight updating unit 37, and a common causal structure node correcting unit 39. Each of these units is a function realized by an instruction from the CPU 301 in Fig. 2 based on a program. In addition, the RAM 303 or the SSD 304 stores a common causal structure DB (Data Base) 21, a user information DB 22, and an advice message DB 23.

[0028] <Explanation of each database> Next, each DB will be explained using FIG. 4 to FIG.

[0029] (Common causal structure DB) Fig. 4 is a conceptual diagram showing nodes and edges of a common causal structure DB as a graph DB, and Fig. 5 is a conceptual diagram of properties of the common causal structure DB as a graph DB.

[0030] As shown in FIG. 4, the common causal structure DB 21 is a graph DB in which the causal relationships are structured, with each node representing a cause that results in the user's physical ailment. Here, the node k(m,n) is the nth descendant (second and subsequent nodes) node from the parent (first level) node 1(0,0) at a distance of m. Each edge has a weight (w) that indicates the relative relationship between multiple nodes (results and multiple causes). node k(from-to, from-to)) are associated. The origin of the arrow is the "cause" and the tip of the arrow is the "result." This weight is a predetermined value that is commonly applied to each user.

[0031] Also, as shown in Fig. 5, information about properties is associated and managed for each node ID. In this case, the information about properties includes physical ailments, causes of ailments, acquisition form of user information, question messages (or sensing items), determination results (Ans) of answers (or sensing data), and thresholds for determining ailments, which are associated and managed. Note that the node ID is an example of node identification information for identifying a node.

[0032] "Illness" indicates a physical illness and indicates a parent node.

[0033] "Cause of illness" indicates the cause of the result of "illness." For example, "accumulation of mental fatigue" and "accumulation of physical fatigue" are managed as causes of the result of "depressive symptoms."

[0034] "User information acquisition form" refers to a method for acquiring information (data) related to illness from each user. "User input" refers to a method for presenting a question from the illness cause estimation device 30 to the user's communication terminal 50 and acquiring an answer. Other than "user input" refers to a method for periodically acquiring physical measurement data from a wearable device (smart watch, etc.) worn by the user. Examples of other than "user input" include "real space behavior log," "server space behavior log," and "vital data."

[0035] "Real-space behavior logs" are information collected by sensors 40 installed in smartphones and smartwatches, POS (Point of Sales) data, etc., and include the user's number of steps, means of transportation, location information, duration of sitting, purchase history at physical stores, etc. From this real-space behavior log data, behaviors such as daily rhythms, meals, and lifestyle habits are analyzed through numerical analysis.

[0036] The "cyberspace behavior log" is a browsing history or posting history on a website (an application on the communication terminal 50), or a purchase history on an EC (electronic commerce) site, etc. From this cyberspace log data, natural language processing or the like is used to analyze mainly the user's emotions, thoughts, feelings, preferences, values, worries, dissatisfaction, desires, goals, interests, concerns, curiosity, concerns, recognition, understanding, interpretation, and other behaviors.

[0037] "Vital data" refers to measurement data such as heart rate, blood pressure, and blood sugar level measured in real time by a sensor 40 such as a wearable device, and also includes measurement data obtained through health checkups.

[0038] The real-space behavior log, the cyber-space behavior log, and the vital data are examples of "sensing data."

[0039] "Question message" indicates a question message presented to the communication terminal 50 of each user when the user information acquisition form is user input.

[0040] The "sensing item" indicates the type of sensing data acquired by the sensor 40 or the like when the user information acquisition form is other than user input.

[0041] "Answer or data judgment" indicates the criteria for judging the answer or sensing data obtained from the user. For example, the criteria may indicate that the answer or sensing data result is evaluated on a five-point scale.

[0042] The "threshold" indicates a threshold for determining whether a person is unwell based on the results of the response or sensing data. For example, if the response or sensing data result is 4 or higher on a 5-point scale, it is determined to be a depressive symptom.

[0043] (User information DB) Fig. 6 is a conceptual diagram showing the nodes and edges of a user information DB as a graph DB. In the user information DB 22, in a graph structure similar to that of the common causal structure DB 21 in Fig. 4, each node is associated with a determination result (Ans) based on response data from the user or sensing data, and each edge is associated with a cause score (Fscore) for management. The determination result (Ans) and the cause score (Fscore) will be explained in detail later.

[0044] (Advice message DB) Fig. 7 is a conceptual diagram of the advice message DB. As shown in Fig. 7, information about advice is managed in association with each node ID. Each node is the same as in Fig. 5. In this case, the information about advice is managed in association with messages indicating malfunctions, causes of malfunctions, and countermeasures. "Malfunctions" and "causes of malfunctions" are the same as in the common causal structure DB in Fig. 5.

[0045] The "message indicating a method of dealing with the problem" is a message indicating a method of dealing with the problem that should be presented to a specified user when the illness cause estimation device 30 estimates the cause (main cause) of the physical illness of the specified user based on the cause score (Fscore).

[0046] (Each functional configuration) Next, the functional configuration of the malfunction factor estimation device 30 will be described.

[0047] When the user state acquisition unit 31 receives a request from the user illness cause search unit 33 to present a question about a specific node and acquire an answer, it displays the question and options on the user's communication terminal 50. Then, it sends data of the user's answer to the user illness cause search unit.

[0048] In addition, when the user status acquisition unit 31 receives a request for periodic sensing from the user illness cause search unit 33, it acquires sensing data from the sensor 40 etc. according to the user information acquisition format (real space behavior log, cyber space behavior log) (see Figure 5) and sends it to the user illness cause search unit 33.

[0049] When the illness cause estimation device 30 senses the user's condition and determines the user's illness state, the user condition acquisition unit 31 uses the numerical values ​​of the acquired sensing data and reference values ​​(e.g., average value, standard deviation) to convert the data into a five-level evaluation as follows:

[0050] 5...less than average -2σ 4...Less than (average -0.5σ) to (average -2σ) or more 3…Average within ±0.5σ 2...greater than (average + 0.5σ) to (average + 2σ) or less 1...greater than the mean -2σ As a result, for example, when the user status acquisition unit 31 acquires sensing data for "sleep time" using a real-space behavior log, it outputs a result of "2" on the above five-point scale using data indicating that the user's sleep time is 7.5 hours, "average sleep time for Japanese people: 6 hours," and "standard deviation: 2 hours."

[0051] The user state acquisition unit 31 also sends the common causal structure weight update unit 37 data at the same time as sending data (answer data or sensing data) to the user illness cause search unit 32 .

[0052] The user illness cause search unit 33 estimates the cause of the physical illness of the specified user by applying data on answers from the specified user to each specified question or each illness judgment result based on sensing data related to the specified user to each node corresponding to each specified question and calculating each cause score related to the illness.

[0053] Specifically, the user illness cause searching unit 33 calculates the judgment result (Ans) of illness for each node based on the data (answer data or sensing data) received from the user state acquiring unit 31, and calculates a cause score (Fscore) using the weights managed in the common causal structure DB 21. Then, the user illness cause searching unit 33 manages the judgment result (Ans) in association with each node in the user information DB 22, and manages the cause score (Fscore) in association with each edge in the user information DB 22.

[0054] The cause score is calculated using (Equation 1).

[0055]

number

[0056] Here, a method for calculating a cause score will be described with reference to Fig. 8. Fig. 8 is a diagram showing a method for calculating a cause score.

[0057] For example, as shown in Figure 8, if node 1 (1,1) of "accumulation of mental fatigue" is the result and node 1 (2,1) of "poor human relationships" is the cause, a weight w1 (2-1,1-1) = 0.3 is previously managed in the common causal structure DB 21. Here, the user's illness cause searching unit 33 calculates Ans(node1 (2,1)) = 4 as the physical illness judgment result based on the data of the user's answers or the user's sensing data from the user state acquisition unit 31. Then, the user's illness cause searching unit 33 calculates Fscore1 (2-1,1-1) = w1 (2-1,1-1) · Ans(node1 (2,1)) = 0.3 * 4 = 1.2 by multiplying the predetermined weight (w1) by the judgment result (Ans) of the predetermined user's illness.

[0058] The processing of the user's illness cause searching unit 33 will be explained in detail later.

[0059] When the advice presenting unit 35 receives a request from the user state acquiring unit 31 to present a node ID (top N factor scores) and a message indicating a solution, the advice presenting unit 35 searches the advice message DB 23 using this node ID as a search key to read out the data of the corresponding "message indicating a solution." Then, the advice presenting unit 35 displays the message indicating the solution on the communication terminal 50. Note that the message sent to the communication terminal 50 may display a self-efficacy confirmation screen to which a comment is added asking the user whether they think they can improve their life in accordance with the advice (see FIG. 11).

[0060] The common causal structure weight update unit 37 can update or otherwise change the weights previously managed in the common causal structure DB 21 using response data or sensing data related to the user's recent illness. For this reason, the common causal structure weight update unit 37 holds a number of data items i required for weight update, which specifies the timing for updating the weights. When the common causal structure weight update unit 37 receives data (response data or sensing data) from the user state acquisition unit 31, it stores the data in a temporary storage area, and starts weight update processing when i pieces of From-side Ans(node ​​k(m,n)) and To-side Ans(node ​​k(m,n)) required for calculating the weight w_node k(from-to, from-to) have been accumulated using data from other users or other data from the same user. The weight update processing is as follows.

[0061] S1: The common causal structure weight update unit 37 calculates the correlation coefficient for i samples using the following (Equation 2):

[0062]

number

[0063]

number

[0064]

number

[0065] S2: The common causal structure weight update unit 37 calculates the correlation coefficient using the following (Equation 3).

[0066]

number

[0067]

number

[0068] This allows, for example, if an infectious disease outbreak suddenly occurs and causes a major change in human behavior, the weights previously managed in the common causal structure DB21 or the weights updated by the common causal structure weight update unit 37 can be adjusted based on the opinions of experts.

[0069] The professional or manager may access the advice message DB 23 and add, delete, or modify messages. The professional or manager may also access the user information DB 22 and register information obtained by actually observing users, or modify data that is already managed based on the obtained information.

[0070] [Processing or operation of malfunction factor estimation device] S11: The user's illness cause searching unit 33 refers to the “user information acquisition form” column of the common causal structure DB 21.

[0071] S12: If the user information acquisition form is "user input" in S11, the user illness cause searching unit 33 requests the user state acquiring unit 31 to perform a process of presenting a question about the parent (first level) node k to the user and acquiring an answer from the user. At this time, the user illness cause searching unit 33 reads out the question message in the "question message / sensing item" column managed as "user input" for the parent node k in the common causal structure DB 21 and sends it to the user state acquiring unit 31. This allows the user state acquiring unit 31 to acquire the question content (message) to be presented to the user from the user illness cause searching unit 33. For example, in FIG. 5, the parent node k is shown as node 1 and node 2, and of these, node 1 is managed as the "user information acquisition form" as "user input." Data indicating the "question message" of this node 1, "I have been feeling depressed for more than two weeks," is sent from the user illness cause searching unit 33 to the user state acquiring unit 31. The user state acquisition unit 31 then sends a question message to the user's communication terminal 50 asking, "Have you been feeling depressed for more than two weeks?" and acquires an answer from the communication terminal 50. Since the "answer / data judgment" is a five-point rating, the answer is acquired based on options such as, for example, "Strongly feel that way (5), Feel that way (4), Don't know (3), Don't feel that way (2), Don't feel that way at all (1)." The user state acquisition unit 31 transmits the acquired answer data to the user illness cause search unit 33.

[0072] S13: On the other hand, if the user information acquisition form is "other than user input" in S11 above, the user illness cause searching unit 33 requests the user state acquiring unit 31 to periodically sense the contents of the sensing items related to the parent (first level) node k or the descendant (second level or later) node k(m, n). At this time, the user illness cause searching unit 33 reads out the sensing items in the "question message / sensing item" column managed as other than "user input" in the parent node k in the common causal structure DB 21, and sends them to the user state acquiring unit 31. This allows the user state acquiring unit 31 to acquire the sensing items from the user illness cause searching unit 33. For example, in FIG. 5, the parent node k is shown as node 1 and node 2, and of these, node 2 is managed as having a "user information acquisition form" other than "user input." Data indicating the "stress level", which is a "sensing item" of node 2, and the like are transmitted from the user illness cause searching unit 33 to the user state acquiring unit 31. The user state acquiring unit 31 then transmits a request for periodic sensing of the "stress level" to the user's sensor 40, and periodically acquires sensing data indicating the sensing results from the sensor 40. The user state acquiring unit 31 sends the acquired sensing data to the user illness cause searching unit 33. Note that if the user state acquiring unit 31 has already periodically acquired sensing data from the sensor 40, the user state acquiring unit 31 does not need to transmit a request for the contents of the sensing item, such as the "stress level", to the sensor 40.

[0073] S14: The user illness cause searching unit 33 manages the data (answer data or sensing data) acquired from the user state acquiring unit 31 in association with the corresponding node ID in the user information DB 22 (FIG. 6). For example, if a judgment result indicating 4 on a 5-point scale (the higher the number, the worse the condition) is obtained from the answer data or sensing data, the user illness cause searching unit 33 manages the judgment result of Ans(node ​​1)=4 for node 1(0,0) of "depression symptoms" in the user information DB 22, as shown in FIG.

[0074] S15: Next, the user illness cause searching unit 33 searches the user information DB 22 and determines whether the user, who is the subject of the answer or sensing, is physically ill based on the response data from the user or the sensing data related to the parent node k. Specifically, the user illness cause searching unit 33 determines whether the result related to the response data from the user or the sensing data is equal to or greater than a threshold, determining that the user is physically ill if the result is equal to or greater than the threshold, and determining that the user is not physically ill (or normal) if the result is less than the threshold. For example, in FIG. 5, the threshold for "depression symptoms" for node 1 (0,0) is 4, so the user illness cause searching unit 33 determines that the user is ill if the determination result related to the response data from the user or the sensing data is "4" or greater, and determines that the user is not ill if the result is less than "4." If the user is not ill (S15; no ill), the process returns to S11.

[0075] S16: If an illness is detected as a result of S15 above (S15; Illness), as shown in FIG. 10, the user illness cause search unit 33 determines whether the response data or the value of the judgment result related to the sensing data is stored for one or more of the descendant nodes k(m, n) in the user information DB22 (see FIG. 6).

[0076] S17: If the result is stored in one or more nodes (S16; YES), the user illness cause searching unit 33 requests the user state acquiring unit 31 to perform a process of presenting questions to the user regarding a predetermined number of nodes k(m, n) with the highest weights among the descendant nodes k(m, n) whose result values ​​are not stored in the user information DB 22, and acquiring answers from the user. The predetermined number of the highest weights is, for example, three. The reason for limiting the number of questions in this way is to avoid the user feeling uncomfortable if asked too many questions, and not answering at all, or the user giving a random answer. In this case, the user illness cause searching unit 33 reads out the question message from the “Question Message / Sensing Item” column managed as “User Input” for the descendant node k(m, n) in the common causal structure DB 21 (FIG. 5), and sends it to the user state acquiring unit 31. In this way, the user state acquiring unit 31 can acquire the question content (message) to be presented to the user from the user illness cause searching unit 33. For example, if it is determined in S15 above that the user is unwell due to "depressive symptoms," data indicating the top three "question messages" (see FIG. 5) with the highest weights for nodes 1(1,1), 1(1,2), etc., which are descendants of node 1(0,0) of "depressive symptoms" in FIG. 4 and which are nodes 1(1,1), 1(1,2), etc., for which the "user information acquisition format" is managed as "user input" in FIG. 5, are sent from the user unwell cause searching unit 33 to the user state acquiring unit 31. The user state acquiring unit 31 then transmits each question message to the user's communication terminal 50 and acquires data on each answer from the communication terminal 50. The user state acquiring unit 31 then transmits the acquired data on each answer to the user unwell cause searching unit 33.

[0077] S18: On the other hand, if the node k(m, n) is not stored in one or more nodes (if the node k(m, n) is not stored at all) (S16; NO), the user illness cause searching unit 33 requests the user state acquiring unit 31 to perform a process of presenting a question to the user regarding a predetermined number of nodes k(m, n) with the highest weights among all descendant nodes k(m, n) managed in the user information DB 22, and acquiring an answer from the user. In this case, the process in which the user illness cause searching unit 33 reads a question message from the common causal structure DB 21 (FIG. 5) and sends it to the user state acquiring unit 31 is the same as S17 above. Note that S17 and S18 are limited to "user input" because if the question is asked to the user, an answer can be obtained quickly by the user operating the communication terminal 50 without the user having to exercise. However, if the process from S19 onwards is not urgent, data other than "user input" may be obtained by sensing.

[0078] S19: After processing S17 or S18, the user illness cause searching unit 33 manages the determination results for the predetermined number of data (answer data or sensing data) acquired from the user state acquiring unit 31 in association with the corresponding node ID in the user information DB 22 (FIG. 6). For example, as shown in FIG. 6, the determination result for the answer data for node 1 (1,1) of "accumulation of mental fatigue" is "3," and the determination result for the answer data for node 1 (1,2) of "accumulation of physical fatigue" is "2."

[0079] S20: The user's illness cause searching unit 33 calculates the cause score of node k(from-to, from-to) using the weight of the edge and the value of node k(m, n) associated with the edge.

[0080] As described above, the cause score of "poor interpersonal relationships" as a cause of "accumulation of mental fatigue" that is causing user A to feel unwell is 1.2. Similar processing by the user's illness cause searching unit 33 results in a cause score of "lack of sleep" as a cause of "accumulation of mental fatigue" that is causing user A to feel unwell is 0.2, as shown in FIG. 6. Therefore, in the case of user A, it is determined that "poor interpersonal relationships" are more influential than "lack of sleep" as a cause of "accumulation of mental fatigue." By repeating the processing of S20, it is possible to infer a multi-stage structured relationship in which "accumulation of mental fatigue" has a large influence as a cause of the "depressive symptoms" of node 1(0,0), and "poor interpersonal relationships" have a large influence as a cause of this "accumulation of mental fatigue."

[0081] S21: The user's illness cause searching unit 33 requests the advice presenting unit 35 to present the top N messages that should be addressed based on cause scores. In this case, the user's illness cause searching unit 33 transmits the node IDs of the top N nodes based on cause scores to the advice presenting unit 35. For example, the top three cause scores may include the following multi-level content: Bad relationships (node ​​ID: node 1(2,1)) Hypoglycemia due to excessive sugar intake (node ​​ID: node 1(1,n)) - Mainly desk work (node ​​ID: node 1(1,5)) As a result, the advice presenter 35 reads out message data indicating a method of dealing with the problem associated with the node ID from the advice message DB, and transmits the message data to the communication terminal 50. Fig. 11 is a diagram showing an example of a display on the communication terminal. As shown in Fig. 11, a message indicating a method of dealing with the problem is displayed on the communication terminal 50.

[0082] This completes the description of the processing or operation of the malfunction factor estimating device.

[0083] [Effects of the embodiment] As described above, according to this embodiment, the answers from the user to a predetermined question or the user's sensing data are applied to a graph that structures the causal relationships with the causes of the user's physical ailment as nodes, and a cause score for each ailment is calculated. This has the effect of making it possible to estimate the cause of the ailment even when the causes of the physical ailment have a multi-stage structure and the cause (main cause, fundamental cause) of the physical ailment is difficult to identify.

[0084] 〔supplement〕 The present invention is not limited to the above-described embodiment, and may have the following configurations or processes (operations). (1) The malfunction factor estimation device 30 can be realized by a computer and a program, but this program can also be recorded on a (non-transitory) recording medium or provided via the communication network 100. (2) In the above embodiment, a smartphone is shown as an example of a communication terminal 50, but this is not limited to this and may be, for example, a node-type personal computer, a desktop personal computer, a tablet terminal, a smartphone, a smart watch, a car navigation device, a refrigerator, a microwave oven, etc. (3) The CPU 301 may be a single CPU or multiple CPUs. [Explanation of symbols]

[0085] 1. Communication Systems 3 Malfunction factor estimation device 5. Communication terminals 21 Common causal structure DB 22 User Information DB 23 Advice Message DB 31 User status acquisition unit 33 User Malfunction Cause Search Department 35 Advice section 37 Common causal structure weight update unit 39 Common causal structure node correction section

Claims

1. An illness cause estimation device that estimates a cause of a user's physical illness, a common causal structure database as a graph database in which causal relationships are structured, with each node representing a cause of the user's physical discomfort; a user illness cause search unit that estimates the cause of the physical illness of the predetermined user by applying data on answers from the predetermined user to each predetermined question or each illness determination result based on sensing data related to the predetermined user to each node corresponding to the predetermined question and calculating each cause score related to the illness; and the common causal structure database is a graph database in which a causal relationship is structured with the result of the user's physical illness, a first cause that causes the result, and a second cause that causes the first cause as the nodes, and the edges are weights that indicate the relative relationship between the result of the illness and the first cause, and the relative relationship between the first cause and the second cause, the user's illness cause searching unit calculates each cause score based on the determination results applied to each node and the weights associated with each node; Malfunction factor estimation device.

2. The malfunction factor estimating device according to claim 1, the common causal structure database manages, for each node, a form of data acquisition from the user as a property; An illness cause estimating device having a user state acquiring unit that acquires the response data from the specified user or the sensing data related to the specified user based on the acquisition format.

3. The malfunction factor estimating device according to claim 1, An illness cause estimating device having an advice presenting unit that presents a message indicating a method of dealing with the cause of the physical illness of the specified user to the communication terminal of the specified user.

4. 2. The illness cause estimation device according to claim 1, wherein each of the weights managed in the common causal structure database is changed based on the response data from the specified user or the sensing data related to the specified user, or is changed by a specific person.

5. A malfunction factor estimation method executed by a malfunction factor estimation device that estimates a cause of a user's physical malfunction, the illness factor estimating device has a common causal structure database as a graph database in which causal relationships are structured, with each cause of the user's physical illness being represented as a node, and The malfunction factor estimation device data on answers from a predetermined user to each predetermined question or each judgment result of illness based on sensing data related to the predetermined user is assigned to each node corresponding to the predetermined question, and each cause score related to illness is calculated, thereby performing a user illness cause search process to estimate the cause of the physical illness of the predetermined user; the common causal structure database is a graph database in which a causal relationship is structured with the result of the user's physical illness, a first cause that causes the result, and a second cause that causes the first cause as the nodes, and the edges are weights that indicate the relative relationship between the result of the illness and the first cause, and the relative relationship between the first cause and the second cause, The user's illness cause search process includes a process of calculating each cause score based on each determination result applied to each node and each weight associated with each node. Methods for estimating the causes of illness.

6. A program causing a computer to execute the method according to claim 5.

Images