JP2022059547A - Mental improvement support device - Google Patents

Abstract

To specify reasons which are making a subject anxious including reasons that the subject does not even know that the reasons are making the subject anxious, and provide the best advice to the subject.SOLUTION: A mental improvement support device asks a question to a subject or measures the subject, constructs an actual situation model such as experiences of the subject from the past to the present days on the basis of the answer or the result, constructs a mental state model such as reasons of anxiety and presents an advice package accordingly. After that, the device optimizes the model construction and the specification of the advice package from feedback such as an answer to a question or the satisfaction level of the advice package.SELECTED DRAWING: Figure 3

Description

The present invention relates to a mental improvement support device, a mental improvement support system, and a program.

In Patent Document 1, the counselor focuses on cognitive distortion and mood, extracts and analyzes common keywords, feeds back the results, and gives the following advice. This series of work is automated by making it into an algorithm and making it into an application. Counseling is provided by classifying the common keywords of qualitative data called cognitive distortion and feeding back the numerical changes. In addition to recommending action items for each purpose, a system is disclosed that supports cognitive-behavioral therapy by improving motivation and continuity because the results appear numerically and the effects are likely to appear.

Japanese Unexamined Patent Publication No. 2017-146705

The system described in Patent Document 1 focuses on cognitive distortion and mood, analyzes keywords, feeds back the results, and gives the following advice as counseling. However, this is limited to the part of cognitive-behavioral therapy that is part of the counseling of psychological counselors. Not only the subject's cognitive distortion and mood, but also the subject's past, relationships, living environment, hormonal balance, sympathetic parasympathetic nerves, and other physiological conditions are the causes of deterioration of the subject's mental state. Correcting the cognitive distortion of the subject is also effective in improving the subject's mental state, but depending on the subject's situation, advice on specific behavioral guidelines and lifestyle advice is given to predict and convey the subject's future. Other actions, such as giving, may be more suitable than the cognitive-behavioral therapy approach. Not only what a psychological counselor can do, but also the fact that human beings such as career counselors, lifestyle counselors, and work habits counselors are divided into different tasks are integrated, and the causes of deterioration of the mental state of the subject are analyzed from multiple perspectives. It is necessary to select and implement the most suitable method from multifaceted identification, and according to it, cognitive behavioral therapy, presentation of behavioral guidelines, presentation of cases of others, and so on.
In addition, if the known cause and effect that can be understood from the theory and the know-how of the existing human counselor are dropped into the system, and if the theory and the human counselor are wrong, this purpose cannot be achieved. Therefore, it is also necessary to create a unique cause and effect in the system.

Therefore, the present invention incorporates various factors such as the external environment surrounding the subject, external factors, the subject's thinking tendency, and mental structure, and creates a model from which the actual condition of the subject is grasped. And run the advice package.
It can be said that the characteristic means of the present invention are data processing, generation method and ingenuity until various information factors are taken in, used for identifying the cause of deterioration of mental state, and an advice package is presented according to the factors.
In addition, machine learning is inserted and optimization is performed in the model creation and the execution of the advice package according to it.
In addition, in order to reduce the stress of the subject in question presentation and measurement up to model creation and advice package execution, and to gain the trust and trust of the subject, the question presentation method and the question presentation method according to the subject's mental structure model Optimize the measurement method.

By searching for the cause of mental deterioration from not only limited subject information such as cognitive distortion and mood, but also multifaceted subject information, causes other than cognitive distortion can be identified, and only cognitive behavioral therapy. It is possible to implement measures to improve mental status from various methods that are not.
In addition, by optimizing the cause of mental deterioration by machine learning, it is possible to find causes and effects that were not known by the current theory and the know-how of human counselors, and use them for the above goals.
In order to change the approach of this system to the target person from the mental state of the target person,
The target person's trust in this system will increase.

It is a diagram of the flow of asking a question, getting an answer, and building a fact-finding model. It is a diagram of the flow until the advice is presented. It is a diagram of the flow of the fact model construction, the mental state model construction, and the advice package. It is a figure of the evaluation score calculation leading to the identification of the cause of trouble. It is a figure to acquire the necessary fact-finding model element by an additional question. It is a diagram that acquires the necessary actual model elements by additional questions and proceeds to identify the cause of the trouble. It is a figure of the presentation example of a definite question. It is a figure of the presentation example of a selection question. It is a diagram of a specific flow of worries classified into a tree structure and causes of worries.

<Flow of building a real-life model>
As shown in FIG. 1, the actual model element is specified from the actual model element information group according to the question answer and the measurement result, and it is accumulated in the actual model of the subject.
The actual situation model is a mental structure model consisting of the subject's personality tendency, thinking bias, things that are the reason for living, etc., and a reality consisting of living environment, surrounding environment, life experience, human relations, work situation, physiological information, etc. Including world model.

<Identifying the cause of the target person's worries and presenting advice package information for it>
In the trouble information group, trouble cause information group, and advice package information group stored in the information storage unit, the goodness of fit with the actual condition model element of the subject, the trouble, and the evaluation score (evaluation score) based on the trouble cause are calculated. Contains a weighting factor for The evaluation score is calculated by, for example, a weighted sum between the actual condition model and the mental state model of the subject and the weighting coefficients of the trouble information group, the trouble cause information group, and the advice package information group. An example is shown in FIG.

As a result, based on the actual situation model of the target person, the problem information group and the problem cause information group corresponding to the problem and the cause of the problem of the target person, and the advice package information of the contents according to the actual situation model of the target person. The group can be estimated from the one with the highest evaluation score.

From the worries information group, worries cause information group, and advice package information group with high evaluation scores, the worries that the subject has and the worries information group and worries cause information group corresponding to the worries cause, and the actual situation model element of the subject. In order to narrow down or identify the candidates for the package information group, the question is presented to the target person and the answer information of the target person is obtained.

The questions presented to the target person are, for example, a question that presents options for narrowing down the corresponding trouble information group and the trouble cause information group (selection question), and whether the estimated trouble information matches the target person's trouble. A question (confirmed question) for confirming the above is included, and is presented, for example, by the information presentation unit as shown in FIGS. 7 and 8, respectively.

In addition, when identifying the worries information group corresponding to the worries of the subject, a question may be presented to identify the cause of the worries in the subconscious mind that the subject cannot verbalize.

By the above processing, the trouble information and the trouble cause information corresponding to the trouble and the subconscious of the subject are specified, and the advice package information group narrowed down to correspond to the trouble information and the trouble cause information is used to obtain the evaluation score. High advice Select package information and perform the step of presenting it to the target person from the information presentation department. This is shown in FIG.

<Feedback>
Based on the subject's presentation of the question, the subject's answer to it, and the subject's feedback data on the presented advice package information, the subject's actual condition model and mental state model are estimated, and the advice package information to be presented is estimated. Give feedback on each estimator to be used and its contents.

This makes it possible to evaluate, for example, whether the subject's worries and the cause of the worries are properly estimated, or whether appropriate advice package information can be presented to the subject, and based on the evaluation information. The accuracy of each estimator can be improved, and the content of the advice package information can be improved.

Feedback is a component of this system when the parameters of each estimator (for example, weighting factor) are automatically corrected from the feedback data of the subject using a predetermined update formula (learning), and from the feedback data of the subject. Includes the case of manually rewriting (for example, the content of the advice package information) (brush up).

In addition, the feedback is when only the feedback data of the target person is reflected (individual feedback) and when the feedback data of all users of this system including the feedback of the target person is reflected (feedback common to others). including.

<Q & A for acquiring information that has not been obtained yet>
This system, for example, when identifying the subject's worries, the cause of the worries, and the advice package information to be presented, asks a group of questions to extract the actual situation model elements of the subject who are lacking in the identification. Present. The actual situation model and the mental state model are updated from the answer information of the subject to the question group.

Here, "insufficient in identification" means, for example, between the trouble cause information group in the information storage unit and the actual condition model element and the mental state model element of the subject when the subject's trouble cause is specified. This includes the case where the evaluation score (evaluation score) based on the goodness of fit is not high only for a certain trouble cause information, and it cannot be determined that the trouble cause information corresponding to the trouble cause of the subject is that.

Every time the actual condition model and mental state model of the subject are updated, in the calculation of the evaluation score between the trouble information, the trouble cause information, and the advice package information in the information storage department and the subject's trouble, the cause of the trouble, For example, the value of a term with a weighted sum changes, and as a result, each evaluation score is updated. This is shown in FIG.

As a result, from the response information of the subject to the presentation of the question group and the information obtained by the measurement, the subject's worries are identified, the cause of the worries is identified, and the advice package information to be presented is identified while sequentially reflecting the actual situation of the subject. The more the target person uses this system, the more the target person's actual situation is accumulated as information, and the advice package information that is more in line with the target person's actual situation can be specified, or the target person's worries can be estimated and worried more accurately. The cause can be estimated. This is shown in FIG.

In addition, the question presented to the target person at this time is, for example, the content of extracting the actual situation model element that leads to the identification of the most trouble information from the corresponding trouble information candidates when the target person identifies a certain trouble. Select. What is "the actual model element that leads to the identification of the most troubled information"? For example, it is known that the evaluation score of the candidate for troubled information differs greatly depending on A for the actual condition model element A that is not included in the actual condition model of the subject. At times, this actual condition model element A becomes "the actual condition model element that leads to the identification of the most troubled information".

As a result, it is possible to shorten the time and procedure required to identify the worries of the target person, the cause of the worries, and the advice package information to be presented.

<Process of narrowing down trouble information group / advice package information group>
When selecting the advice package information to be presented to the target person, first identify the target person's worries (worry identification step) and then identify the cause of the worries (worry cause identification step). After that, the advice package information that matches the cause of the trouble and the actual situation model element of the target person is selected (advice package information identification step). This process is described in FIGS. 2 and 3.

The worries and causes of worries of the subject, as well as the worries and worries cause information in the information storage unit, are classified into categories (worries category) having a hierarchical structure for each degree of abstraction (shown in FIG. 9). In the trouble identification step and the trouble cause identification step, the trouble category with a high degree of abstraction is gradually classified into a specific trouble category with a low degree of abstraction based on the presentation of the question and the answer of the target person. , Narrow down and identify the worries of the target person and the candidates for the cause of the worries.

The questions presented to the target person in the trouble identification step and the trouble cause identification step are selections that present questions and options for determining the classification destination when the target person's trouble category is further embodied. Includes a question and a definitive question to determine if the estimation is correct for the subject's worries category estimated from the subject's factual and mental state model elements.

Choice questions include, for example, presenting a list of child categories of a category and allowing the subject to select one or more that they think fit their worries.

The definite question is a question and an option for confirming whether the estimation is correct when there is a category of the subject's worries and worries estimated from the worries information group with a high evaluation score and the worries cause information group. Present (eg, "Are you worried about XX?" Yes / No).

Since the evaluation score of the trouble information candidate is high, when the classification to the trouble category of a certain abstraction and the classification candidate can be estimated with sufficient accuracy, a definite question to confirm whether the estimation is correct may be presented. be.

If the subject answers that the estimation is correct for the definite question, the narrowing down is executed. If the estimation is incorrect, the selection question may be presented again in order to narrow down the candidates, or another definite question may be presented again.

If the estimation of the trouble information group by presenting the definite question is correct, it is possible to make the target person think that this system understands the actual situation of the target person or solves the target person's trouble, and as a result, the target person. It is possible to greatly improve the credibility and trust of a person's system.

Each trouble information is associated with a group of advice package information corresponding to it. After identifying the trouble information corresponding to the trouble of the target person by the above processing, the advice package information having a high evaluation score is selected from the advice package information group corresponding to the trouble and presented to the target person.

For example, when the classification to the trouble category with a high degree of abstraction and the classification candidate to the trouble category with a low degree of abstraction can be sufficiently estimated by the definite question, the definite question to confirm the estimation of the classification to the trouble category with a lower degree of abstraction is asked. It should be presented to the target person with priority.

The reliability score of the subject can be greatly improved by making a more specific estimation of the subject's worries.

Since the evaluation score of the trouble information group candidate to be narrowed down when a certain option is selected is low, it may be hidden when it is estimated that the target person is unlikely to be selected. be. On the contrary, there are cases where the options that are presumed to be selected by the subject are presented in descending order. In addition, when it is estimated that there is only one option that the target person is likely to answer in the elective question and it is unlikely that the other option is selected, the selection question is automatically narrowed down and presented. It may be omitted.

By these operations, it is possible to prevent the target person from presenting more options than necessary, and it is possible to reduce the time and effort for the target person to search for his / her own answer from multiple options. Therefore, the target person when using this system. Stress can be reduced.

<About estimation and learning of credit and skepticism>
In order to estimate the trust of the subject in this system, the index of the subject's credit value and suspicion value will be introduced.
When presenting the above question, these indicators are referred to and used as a basis for determining how to present the question.

For example, if the subject's credit value is low or the suspicion value is high, the frequency of presenting the definite question is suppressed, and conversely, if the subject's credit value is high or the suspicion value is low, the definite question is presented. In some cases, you may be asked to present a definite question to increase the frequency of presenting or to classify into a highly abstract trouble category.

This allows questions and measurements to be made while maintaining or improving the credibility of the subject's system.

The credit value and suspicion value of the subject will increase or decrease based on the subject's answer to the presentation of the above question. At this time, the amount of increase / decrease in the credit value / suspicion value differs depending on the mental structure model element of the subject, and this fluctuation amount is also updated every time the mental structure model of the subject is updated. Furthermore, this amount of fluctuation may change depending on the number of times the system has been used and the time it has been used by the subject.

The amount of increase / decrease in the credit value / suspicion value may be determined, for example, by the classification of the subject's mental tendency information.

This monitors changes in the reliability parameters estimated according to the subject's mental tendency and the system usage time, and changes the behavior of the system so as not to reduce or improve the subject's credibility with the system. be able to.

<Others>
For the mental structure element that is a part of the fact-finding model element, the personality tendency may be obtained from the question and answer by, for example, personality diagnosis called BIG5 or AQ or calculation of the autism spectrum index.

Mental structural elements include the axis of life, which is the reason for the subject to live.

Real-world elements include the temperature of the day, which may identify the function of the subject's sympathetic parasympathetic nerves and build a fact-finding model.

The cause of worries includes, for example, the bias of the idea that one must obey the instructions of a person in a higher position and the original experience of childhood that caused the bias.

Claims (20)

Correcting cognitive distortions by changing the causes of worries such as the factual model element information, which is the experience of the subject from the past to the present, and the bias of thinking methods that affect human emotions and ways of thinking, in a realistic and flexible manner. The advice package information for solving or improving the happiness is classified by the mental state model element or the actual condition model element including the trouble and the cause of the trouble, or they are accumulated while having them as attributes with weight information. Information storage department and
The fact-finding model building department that analyzes the answer results to the question group to confirm the actual situation of the person obtained from the subject or the result of measurement to confirm the actual situation of the person and builds the actual situation model.
A mental state model that constructs a mental state model consisting of mental state model elements such as the subject's current worries, their worries, and the causes of worries estimated from the actual situation model, based on the analysis results and the actual situation model by the actual situation model construction unit. With the construction department
Based on the mental state model constructed by the mental state model building unit, the mental state model element, the actual condition model element which is a component of the actual condition model obtained by the actual condition model building unit, and the information storage. An evaluation score is calculated based on the goodness of fit (matching) with the mental state model element, which is an attribute of the advice package accumulated in the department, and the actual condition model element, and the one with the highest score is presented to the subject based on the calculation. The advice package identification department that selects multiple advice packages to be done,
A mental improvement support device including an information presentation unit that presents advice packages selected in the advice package specific unit to the target person in descending order of the evaluation score. Of the advice package information presented by the information presentation unit, a weighting coefficient is assigned according to the degree of the positive recovery in association with the advice package information that leads to the positive recovery of the mental state (trouble relief or satisfaction) for the subject. Equipped with a confidence feedback generator to generate tuned confidence parameters
The advice package specifying unit is an attribute corresponding to the evaluation score corresponding to the mental state model element and the actual condition model element or each element at the time of calculating the evaluation score based on the reliability parameter generated by the parameter generation unit. Change the weight information of
The information presenting unit updates the advice package information associated with the trouble and the cause of the trouble and the reliability parameter corresponding to the actual condition model element and the mental state model element and the actual condition model element to the information storage unit. The mental improvement support device according to claim 1, wherein the mental improvement support device is characterized in that it is memorized. In the actual situation model construction unit, in order to acquire information that is necessary when constructing the actual situation model but has not yet been obtained, a question is asked and an answer is obtained or a measurement is obtained and a result is obtained. Identify the actual condition model element that applies to, and add or rewrite it to the actual condition model of the subject. , Ask a question and get an answer or measure to get a result, identify the actual model element applicable to the target person by those answers or results, and add or rewrite it to the actual situation model of the target person. In order to obtain the factual model element and mental state model element that are necessary but not yet obtained, ask a question, obtain an answer, or measure and obtain a result, and the actual situation that applies to the subject by those answers or results. In order to identify the model element and the mental state model element, and to determine the actual condition model, the mental state model element, the mental model element, and the advice package to be added to or rewritten to the actual condition model and the mental state model of the subject, the above-mentioned actual condition model building unit and the mental state model. A request characterized by asking a question, obtaining an answer, or measuring and obtaining a result in the construction department and the advice package specific department, and determining the actual situation model element, the mental model element, and the advice package based on the answer or the result. The mental improvement support device according to Item 1. In the actual situation model construction unit, among the actual situation model elements composed of the real world model and the mental structure model, the one belonging to the real world model is the real world model element, and the one belonging to the mental structure model is the spirit. The real world model element, which is a structural model element, consists of living environment, surrounding environment, life experience, human relations, work situation, physiological information, etc.
The mental structure model element consists of human mental structure, peculiar way of thinking and bias. The real world model element and the mental structure model element are stored in the information storage unit as real world model element information and mental structure model element information. The mental improvement support device according to claim 1, wherein the mental improvement support device is characterized by being accumulated. In the mental state model construction unit, the constructed mental state model consists of mental state model elements, and the mental state model elements consist of worries and causes of worries. Become,
The cause of anxiety consists of the bias of thinking methods and the combination of the real world that caused a certain anxiety.
The mental improvement support device according to claim 1, wherein the worries and the causes of worries are stored in the information storage unit as worries information and worries cause information. In the fact-finding model building section and the information storage section,
Depending on the answer in the question and the result in the measurement, the answer and the result itself may construct the actual condition model of the subject as the actual condition model element of the subject. The actual condition model element information weights the answer and the measurement result of a specific question. Have as an attribute with
An evaluation score is calculated based on the goodness of fit (matching) between the answer to the question and the result in the measurement and the specific answer and the measurement result, which are the attributes of the actual model element, and the actual model element is specified based on the evaluation score. May be added or rewritten to the model At this time, questions and measurements are asked and measured to obtain the answer or result of the question or measurement that is the attribute of the fact model element necessary to identify a certain fact model element. The mental improvement support device according to claim 1, wherein the evaluation score changes depending on the result, the score of a certain actual model element becomes higher than that of others, and the identification of the actual model element may proceed. In the mental structure model construction unit and the information storage unit,
Trouble information has a specific fact-finding model element as an attribute with weight information.
The evaluation score is calculated based on the goodness of fit (matching) between the actual condition model element that constitutes the actual condition model of the target person and the actual condition model element that is the attribute of the trouble, and the trouble is identified from the one with the highest score based on the goodness of fit (matching). Adding or rewriting to a person's worries At this time, in order to obtain the actual model elements that are the attributes of the worries, which are necessary to identify a certain worries, questions and measurements are asked, and the evaluation score changes depending on the results. The mental improvement support device according to claim 1, wherein the score of a certain trouble may be higher than that of another, and the identification of the trouble may proceed. In the mental structure model construction unit and the information storage unit,
The trouble cause information has a specific trouble and the actual situation model element as an attribute with weight information.
An evaluation score is calculated based on the goodness of fit (matching) between the actual condition model element that constitutes the subject's anxiety and the actual condition model, and the anxiety and the actual condition model element that is the attribute of the cause of the annoyance, and the one with the highest score is calculated based on the goodness of fit (matching). Identify the cause of the worries and add or rewrite it to the cause of the worries of the subject. The mental improvement support device according to claim 1, wherein the evaluation score changes depending on the results, the score of a certain trouble becomes higher than that of the other, and the cause of the trouble is identified. In the advice package identification section and the information storage section,
In order to obtain the mental structure model element and the factual model element, which are the attributes of the advice package, which are necessary to identify an advice package, questions and measurements are made, and the evaluation score changes depending on the results. The mental improvement support device according to claim 1, wherein the score of the advice package is higher than that of others, and the identification of the advice package is advanced. When an answer or result is obtained in a question or measurement in the fact-finding model building unit, the mental structure model building unit, or the advice package specifying unit, depending on the answer or result,
Attribute weight information of certain factual model element information, mental structure model element information, advice package information related to the question or measurement changes, or certain actual condition model element information, mental structure model related to the question or measurement It is characterized in that element information and advice package information are specified, or machine learning is performed by rewriting certain actual model element information, mental structure model element information, and advice package information itself related to the question or measurement. The mental improvement support device according to claim 1. According to the reliability parameters generated in the trust feedback generation unit in the actual situation model construction unit, the mental structure model construction unit, the advice package identification unit, and the trust feedback generation unit.
Attribute weight information of certain factual model element information, mental structure model element information, advice package information related to the question or measurement changes, or certain actual condition model element information, mental structure model related to the question or measurement It is characterized in that element information and advice package information are specified, or machine learning is performed by rewriting certain actual model element information, mental structure model element information, and advice package information itself related to the question or measurement. The mental improvement support device according to claim 1. The mental improvement device according to claim 1 is
The mental improvement support device according to claim 1, wherein each user has an account and they are managed by a unified server. In the fact-finding model building unit, the mental structure model building unit, the advice package specifying unit, and the trust feedback generation unit.
The machine learning according to claims 2, 10 and 11 is a case where machine learning is performed only within the user's individual account and a case where machine learning is performed in common in all user accounts. The mental improvement support device according to claim 1, wherein the mental improvement support device may be reflected in an account of another user group having a physical condition model and a mental state model close to the above. In the fact-finding model building unit, the mental structure model building unit, the advice package specifying unit, and the trust feedback generation unit.
An evaluation score based on the goodness of fit (matching) between a user's actual condition model element, mental state model element and another user's actual condition model element, and mental state model element is calculated and held as the degree of similarity between users.
The mental improvement support device according to claim 1. In the fact-finding model building unit, the mental structure model building unit, the advice package specifying unit, and the trust feedback generation unit.
If the user similarity between one user and another is higher than a certain value,
The mental improvement support device according to claim 1, wherein machine learning in one user may be reflected in another user's account. In the actual situation model construction unit, the mental structure model construction unit, the advice package identification unit, the trust feedback generation unit, and the information storage unit.
The actual situation model element and the mental structure model element are specified by the question and the answer to it, the measurement and the result, and at that time, the information storage unit contains the absolute time information in which the element is specified and the book when the element is specified. The number of times the support device has been used is accumulated.
Some actual condition model elements and mental state model elements change or disappear by comparing the accumulated absolute time and the number of times of use with the absolute time and the number of times of use at a certain time.
The mental improvement support device according to claim 1. In the mental state model construction unit,
By obtaining multiple questions and answers, measurements and their results, we will obtain a large number of factual model elements or mental state model elements.
The more they are obtained, the higher the evaluation value of a particular mental state model element will be compared to other mental state model elements.
Along with that, the mental state model element may have the category attribute information of the tree structure, and the category attribute information of the tree structure is constructed so as to be from a wider category to a more detailed category.
Category attributes are narrowed down to more detailed categories by getting answers to specific questions.
When the categories are narrowed down to more detailed categories, the evaluation value of a specific mental state model element becomes higher than that of other mental state model elements.
When the evaluation value of the mental state model element and the numerical value calculated by the evaluation value of the mental state model element below a certain category exceed a certain value, the truth is asked whether the target is facing the mental state model element. Or it may present a definitive question, which is a question that confirms that a particular category contains the mental state model element that the subject is facing.
If true is selected in a question asking whether an object is facing a certain mental state model element, the mental state model element is specified.
If false is selected, the rating of that particular mental state model element is lowered, and by asking more questions and getting answers, the rating of some mental state model elements other than that particular mental state model element is evaluated. The value goes up further.
The tree structure of a category may have a structure that always arrives at a specific mental state model element by repeatedly getting questions and answers by following it.
The mental improvement support device according to claim 1. In the mental state model construction unit,
When you can narrow down to a detailed category by getting the answer to the question about the category attribute of the tree structure,
Each of the answer options to a question may be in a one-step detailed category.
At that time, a one-step detailed category is presented as an option for multiple questions,
The presentation order is determined by the average or total of the estimated values of the mental state model elements existing under the detailed category corresponding to each option, and the category estimated values obtained by other calculation formulas.
The mental improvement support device according to claim 1. In the mental state model construction unit,
For questions that follow tree-structured categories, whether or not to ask a definitive question is determined by whether the category estimates or mental state model element estimates associated with the question exceed a certain definite question value.
The mental improvement support device according to claim 1. In the mental state model construction unit,
The definite question value is determined by the subject's credit suspicion score.
The subject's credit suspicion score shall be calculated by giving specific weights to specific fact-finding model elements of the subject's mental structure model, the number of times the device has been used, the number of false definite questions, and the number of true questions. It is a score determined by.
In addition, if there are multiple definite questions that exceed the definite question value, the definite question belonging to the more detailed one with a tree structure is preferentially presented, or each definite question has an attribute related to the credit skepticism score. , Which definite question is presented is determined according to its attributes.
The mental improvement support device according to claim 1.

Images