Psychological scale recommendation method and device based on artificial intelligence, and storage medium

Abstract

The present application relates to the field of artificial intelligence technology, and discloses a psychological scale recommendation method and device based on artificial intelligence, a storage medium, and a computer device, the method comprising: obtaining the initial input information of the inquiring user and the inquiry dialogue information, wherein the inquiry dialogue information includes the information that has been asked and the information that has been answered; determining the inquiry feature information of the inquiring user based on the initial input information and the inquiry dialogue information; inputting the inquiry feature information into the psychological scale recommendation model, and recommending the psychological scale to the inquiring user based on the obtained psychological scale recommendation data. The present application solves the problem of inaccurate selection when the user selects the psychological scale by himself, and solves the problem of low efficiency and high labor cost caused by the psychological scale recommended by a psychologist, thereby improving the accuracy of the recommendation of the psychological scale and reducing the labor cost.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a psychological scale recommendation method and device based on artificial intelligence, a storage medium, and a computer device. Background Art

[0002] With the continuous improvement of social level, more and more people begin to pay attention to mental health, and more and more people choose to get help with psychological problems through psychological counseling. At present, with the development of Internet technology, online remote psychological counseling through the Internet has become more and more accepted by people due to its convenience and speed, and online psychological counseling has become very common.

[0003] In the process of psychological counseling, psychological scales are a very important method for measuring the psychological problems of the inquirer and the severity of the problems. At present, there are two main ways for online psychological counseling services to provide psychological scales. One is to provide all supported psychological scales to the inquirer, and the inquirer can choose a psychological scale for evaluation based on his or her own situation; the other is that the inquirer first talks with the psychological counselor online, and then the psychological counselor recommends a psychological scale. However, both methods have shortcomings. On the one hand, there is the problem of the accuracy of the psychological scale selection, and on the other hand, there is a shortage of online psychological counselors and low consultation efficiency. Summary of the invention

[0004] In view of this, the present application provides a psychological scale recommendation method and device, storage medium, and computer equipment based on artificial intelligence, which improves the accuracy of psychological scale recommendations and reduces labor costs.

[0005] According to one aspect of the present application, a method for recommending a psychological scale based on artificial intelligence is provided, the method comprising:

[0006] Obtaining initial input information of the inquiring user and inquiry dialogue information, wherein the inquiry dialogue information includes the inquired information and the answered information;

[0007] Determining the inquiry characteristic information of the inquiring user according to the initial input information and the inquiry dialogue information;

[0008] The inquiry feature information is input into a psychological scale recommendation model, and based on the obtained psychological scale recommendation data, a psychological scale is recommended to the inquiry user.

[0009] Optionally, the initial input information includes user characteristics and query keywords, the user characteristics include the age and gender of the inquiring user, and the query keywords are obtained by matching the query description information of the inquiring user with preset keywords;

[0010] Determining the inquiry characteristic information of the inquiring user based on the initial input information and the inquiry dialogue information specifically includes:

[0011] Generate a name query feature vector, a role query feature vector, and a location query feature vector in sequence according to the initial input information and the query dialogue information;

[0012] The name query feature vector, the role query feature vector, and the position query feature vector are accumulated according to the position of each component in the vector to determine the query feature information.

[0013] Optionally, the psychological scale recommendation model is used to predict the matching degree between the inquiring user and each psychological scale;

[0014] The recommending a psychological scale to the inquiring user based on the obtained psychological scale recommendation data specifically includes:

[0015] Determine whether there is any psychological scale in the psychological scale recommendation data whose matching degree with the inquiring user is higher than a preset matching degree threshold;

[0016] If so, the psychological scale that is higher than the preset matching degree threshold is used as the target psychological scale, and the target psychological scale is recommended to the inquiring user.

[0017] Optionally, after determining whether there is any psychological scale in the psychological scale recommendation data whose matching degree with the inquiring user is higher than a preset matching degree threshold, the method further includes:

[0018] If not, input the query feature information into the query information recommendation model to obtain the information to be queried, output the information to be queried to the queried user, and receive the answer information of the queried user to the information to be queried;

[0019] The information to be inquired and the corresponding answer information are added to the inquiry dialogue information as new inquired information and new answered information, and the inquiry feature information of the inquiring user is returned to be re-determined.

[0020] Optionally, the components in the name query feature vector are, in sequence, a user feature component, at least one query keyword component, and at least one group of query dialogue information components, wherein the user feature component includes an age value component and a gender value component, and the query dialogue information component includes a asked information component and an answered information component;

[0021] Each component in the role inquiry feature vector is a feature component symbol corresponding to each component in the name inquiry feature vector, and the role inquiry feature vector includes a user feature symbol, an inquired information symbol, and an answered information symbol;

[0022] Each component in the location query feature vector represents the generation order of each component in the name query feature vector.

[0023] Optionally, the query information recommendation model includes an input representation layer, a network layer and an output layer, wherein the input representation layer includes a name embedding layer, a role embedding layer and a position embedding layer, and the input representation layer is used to accumulate the feature vectors of the name embedding layer, the role embedding layer and the position embedding layer according to the position of each component in the vector;

[0024] The psychological scale recommendation model includes the input representation layer, the network layer, the fully connected layer and the classification layer.

[0025] Optionally, the query information recommendation model and the psychological scale recommendation model have the same network layer parameters, and the training step of the network layer parameters includes:

[0026] According to the sample initial input information of the sample inquirer and the sample inquiry dialogue information, the inquiry information recommendation model is trained to determine the first network layer parameters;

[0027] Using the first network layer parameters as the network layer parameters of the psychological scale recommendation model, training the psychological scale recommendation model according to the sample initial input information of the sample inquirer, the sample inquiry dialogue information and the sample psychological scale category, and determining the second network layer parameters of the network layer;

[0028] The second network layer parameters are used as network layer parameters of the query information recommendation model, and the query information recommendation model and the psychological scale recommendation model are cyclically trained until a training end condition is met.

[0029] According to another aspect of the present application, a psychological scale recommendation device based on artificial intelligence is provided, the device comprising:

[0030] An information acquisition module, used to acquire the initial input information of the inquiring user and the inquiry dialogue information, wherein the inquiry dialogue information includes the inquired information and the answered information;

[0031] A feature determination module, used to determine the inquiry feature information of the inquiring user based on the initial input information and the inquiry dialogue information;

[0032] The recommendation module is used to input the inquiry feature information into the psychological scale recommendation model, and recommend the psychological scale to the inquiring user based on the obtained psychological scale recommendation data.

[0033] Optionally, the initial input information includes user characteristics and query keywords, the user characteristics include the age and gender of the inquiring user, and the query keywords are obtained by matching the query description information of the inquiring user with preset keywords;

[0034] The feature determination module is also used to: generate a name query feature vector, a role query feature vector and a location query feature vector in sequence according to the initial input information and the query dialogue information; accumulate the name query feature vector, the role query feature vector and the location query feature vector according to the position of each component in the vector to determine the query feature information.

[0035] Optionally, the psychological scale recommendation model is used to predict the matching degree between the inquiring user and each psychological scale; and the recommendation module is further used to:

[0036] Determine whether there is any psychological scale in the psychological scale recommendation data whose matching degree with the inquiring user is higher than a preset matching degree threshold;

[0037] If so, the psychological scale that is higher than the preset matching degree threshold is used as the target psychological scale, and the target psychological scale is recommended to the inquiring user.

[0038] Optionally, the recommendation module is further used to:

[0039] After determining whether there is any psychological scale in the psychological scale recommendation data and the inquiring user has a matching degree higher than a preset matching degree threshold, if not, inputting the inquiry feature information into the inquiry information recommendation model to obtain the information to be inquired, outputting the information to be inquired to the inquiring user, and receiving the inquiring user's answer information to the information to be inquired;

[0040] The information to be inquired and the corresponding answer information are added to the inquiry dialogue information as new inquired information and new answered information, and the inquiry feature information of the inquiring user is returned to be re-determined.

[0041] Optionally, the components in the name query feature vector are, in sequence, a user feature component, at least one query keyword component, and at least one group of query dialogue information components, wherein the user feature component includes an age value component and a gender value component, and the query dialogue information component includes a asked information component and an answered information component;

[0042] Each component in the role inquiry feature vector is a feature component symbol corresponding to each component in the name inquiry feature vector, and the role inquiry feature vector includes a user feature symbol, an inquired information symbol, and an answered information symbol;

[0043] Each component in the location query feature vector represents the generation order of each component in the name query feature vector.

[0044] Optionally, the query information recommendation model includes an input representation layer, a network layer and an output layer, wherein the input representation layer includes a name embedding layer, a role embedding layer and a position embedding layer, and the input representation layer is used to accumulate the feature vectors of the name embedding layer, the role embedding layer and the position embedding layer according to the position of each component in the vector;

[0045] The psychological scale recommendation model includes the input representation layer, the network layer, the fully connected layer and the classification layer.

[0046] Optionally, the query information recommendation model and the psychological scale recommendation model have the same network layer parameters, and the device further includes: a model training module, which is used to:

[0047] According to the sample initial input information of the sample inquirer and the sample inquiry dialogue information, the inquiry information recommendation model is trained to determine the first network layer parameters;

[0048] Using the first network layer parameters as the network layer parameters of the psychological scale recommendation model, training the psychological scale recommendation model according to the sample initial input information of the sample inquirer, the sample inquiry dialogue information and the sample psychological scale category, and determining the second network layer parameters of the network layer;

[0049] The second network layer parameters are used as network layer parameters of the query information recommendation model, and the query information recommendation model and the psychological scale recommendation model are cyclically trained until a training end condition is met.

[0050] According to another aspect of the present application, a storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned artificial intelligence-based psychological scale recommendation method is implemented.

[0051] According to another aspect of the present application, a computer device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the above-mentioned artificial intelligence-based psychological scale recommendation method when executing the program.

[0052] Through the above-mentioned technical scheme, the present application provides a psychological scale recommendation method and device based on artificial intelligence, storage medium, and computer equipment, which generates the inquiry feature information of the inquiring user based on the initial input information and inquiry dialogue information of the inquiring user, and pre-constructs a psychological scale recommendation model, and uses the model to automatically recommend psychological scales to users according to the inquiry feature information, thereby solving the problem of inaccurate selection of psychological scales by users themselves, and solving the problem of low efficiency and high labor cost caused by psychologists recommending psychological scales, thereby improving the accuracy of psychological scale recommendations and reducing labor costs.

[0053] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0055] Figure 1 A schematic diagram of a process of a psychological scale recommendation method based on artificial intelligence provided in an embodiment of the present application is shown;

[0056] Figure 2 A schematic diagram of the structure of a query information recommendation model provided in an embodiment of the present application is shown;

[0057] Figure 3 A schematic diagram of the structure of a psychological scale recommendation device based on artificial intelligence provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0058] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0059] In this embodiment, a psychological scale recommendation method based on artificial intelligence is provided. Figure 1 As shown, the method includes:

[0060] Step 101: Acquire the initial input information of the inquiring user and the inquiry dialogue information, wherein the inquiry dialogue information includes the inquired information and the answered information.

[0061] The embodiments of the present application can be used to recommend a psychological scale to an online inquiry user, so as to evaluate the psychological problems of the inquiry user and the severity of the psychological problems with the help of the measurement results of the psychological scale.

[0062] First, the initial input information and inquiry dialogue information of the inquiring user are obtained. The initial input information may be basic information actively input by the inquiring user without inquiring the inquiring user, and may include identity information such as age, gender, and occupation input by the inquiring user, and may also include description information about the inquiring user's own psychological problems input by the inquiring user. The inquiry dialogue information may be questions asked to the inquiring user and answers given by the inquiring user to the questions asked, that is, asked information and answered information.

[0063] Step 102: determining the inquiry feature information of the inquiring user according to the initial input information and the inquiry dialogue information.

[0064] Next, according to the input format of the psychological scale recommendation model, feature extraction is performed on the initial input information and the inquiry dialogue information to obtain the inquiry feature information of the inquiring user.

[0065] In the embodiment of the present application, optionally, the initial input information includes user characteristics and query keywords, the user characteristics include the age and gender of the inquiring user, and the query keywords are obtained by matching the query description information of the inquiring user with preset keywords; step 102 specifically includes:

[0066] Step 102-1, based on the initial input information and the query dialogue information, sequentially generate a name query feature vector, a role query feature vector, and a location query feature vector;

[0067] Step 102-2: Accumulate the name query feature vector, the role query feature vector, and the position query feature vector according to the position of each component in the vector to determine the query feature information.

[0068] In the above embodiment, gender and age can be extracted from the basic information entered by the inquiring user to describe the user characteristics of the inquiring user, and the description information of the psychological problems entered by the inquiring user is matched with keywords according to the pre-built keyword table containing preset keywords, and the query keywords that hit the keyword table are obtained, for example, the query keyword can be "pressure". Thus, based on user characteristics, query keywords, asked information and answered information, feature vectors are generated according to multiple dimensions, including name query feature vectors, role query feature vectors and location query feature vectors.

[0069] Among them, the components in the name query feature vector are user feature components, at least one query keyword component and at least one group of query dialogue information components in sequence, wherein the user feature components include age value components and gender value components, and the query dialogue information components include asked information components and answered information components; the components in the role query feature vector are feature component symbols corresponding to the components in the name query feature vector, and the role query feature vector includes user feature symbols, asked information symbols and answered information symbols; the components in the location query feature vector represent the generation order of the components in the name query feature vector.

[0070] In the embodiment of the present application, the name query feature vector is composed of a user feature component, a query keyword component, and a query dialogue information component. For example, a certain inquiring user has an age value component of C1 and a gender value component of C2. There are two query information keywords, and the query keyword components are K1 and K2 respectively. Two questions have been asked to the inquiring user, and the corresponding queried information components are D1 and D2 respectively. The inquiring user answers these two questions, and the answer information component corresponding to D1 is P1, and the answer information components corresponding to D2 are P2 and P3. Then the name query feature vector is [C1, C2, K1, K2, D1, P1, D2, P2, P3].

[0071] Each component in the role inquiry feature vector is the category symbol of each component in the name inquiry feature vector, that is, the feature component symbol. After determining each component in the name inquiry feature vector, the component at the corresponding position in the role inquiry feature vector can be determined according to the category of each component in the name inquiry feature vector. For example, the feature component symbol of the user feature component is R1, the feature component symbol of the query keyword component is R2, the feature component symbol of the queried information component is R3, and the feature component symbol of the answered information component is R4. Then the role inquiry feature vector corresponding to the above name inquiry feature vector is [R1, R1, R2, R2, R3, R4, R3, R4, R4].

[0072] Each component in the location query feature vector is the order in which each component in the name query feature vector is generated. For example, user features and query keywords are extracted based on the initial input information of the query user. The generation order of the initial input information is 1, represented by S0, and then the query user is asked questions. The generation order of the first question is 2, represented by S1, the generation order of the answer to the first question is 3, represented by S2, the generation order of the second question is 4, represented by S3, and the generation order of the two answers to the second question is 5 and 6, represented by S4 and S5, respectively. Then the location query feature vector corresponding to the above name query feature vector is [R1, R1, R2, R2, R3, R4, R3, R4, R4].

[0073] In the above manner, a name query feature vector, a role query feature vector, and a position query feature vector with the same length can be obtained, and further accumulation is performed according to the position of each component in the vector to determine the query feature information.

[0074] Step 103: input the inquiry feature information into a psychological scale recommendation model, and recommend a psychological scale to the inquiring user based on the obtained psychological scale recommendation data.

[0075] Finally, the inquiry feature information is used as input, and the psychological scale of the inquiring user is predicted through the pre-built and trained psychological scale recommendation model. According to the output result of the psychological scale recommendation model (i.e., psychological scale recommendation data), a suitable psychological scale is recommended to the inquiring user.

[0076] By applying the technical solution of this embodiment, the inquiry feature information of the inquiring user is generated based on the initial input information of the inquiring user and the inquiry dialogue information, and by pre-building a psychological scale recommendation model, the model is used to automatically recommend psychological scales to users according to the inquiry feature information, thereby solving the problem of inaccurate selection of psychological scales by users themselves and the problem of low efficiency and high labor cost caused by psychologists recommending psychological scales, thereby improving the accuracy of psychological scale recommendations and reducing labor costs.

[0077] In the embodiment of the present application, optionally, the psychological scale recommendation model is used to predict the matching degree between the inquiring user and each psychological scale; in step 103, "recommending a psychological scale to the inquiring user based on the obtained psychological scale recommendation data" specifically includes:

[0078] Step 103 - 1 , determining whether there is any psychological scale in the psychological scale recommendation data whose matching degree with the inquiring user is higher than a preset matching degree threshold.

[0079] Step 103 - 2 : If there is a psychological scale that is higher than the preset matching degree threshold, the psychological scale is used as a target psychological scale, and the target psychological scale is recommended to the inquiring user.

[0080] Step 103-3, if it does not exist, input the inquiry feature information into the inquiry information recommendation model to obtain the information to be asked, and output the information to be asked to the inquiring user, and receive the inquiring user's answer information to the information to be asked; add the information to be asked and its corresponding answer information as new asked information and new answered information to the inquiry dialogue information, and return to step 102.

[0081] In this embodiment, the psychological scale recommendation model can predict the matching degree between the inquiring user and each psychological scale based on the inquiry feature information. After the psychological scale recommendation model performs the prediction, it is first determined whether there is a psychological scale with a matching degree greater than a preset matching degree threshold among the psychological scales based on the prediction result. If there is a psychological scale with a matching degree greater than the preset matching degree threshold, it means that the psychological scale that the inquiring user should measure has been sufficiently evaluated based on the current initial input information and inquiry dialogue information, and the psychological scale can be recommended to the inquiring user at present, and the psychological scale with a matching degree greater than the preset matching degree threshold is recommended to the inquiring user as the target psychological scale.

[0082] However, if there is no psychological scale with a matching degree greater than the preset matching degree threshold, it means that the current initial input information and inquiry dialogue information are not sufficient to evaluate the psychological scale that the inquiring user should measure, and further understanding of the psychological problems of the inquiring user is needed to determine which psychological scale should be measured. At this time, you can continue to ask questions to the inquiring user so as to further understand the user's psychological problems through the user's answers.

[0083] In the present application, the next inquiry question for the inquiring user can also be determined by a pre-built and trained inquiry information recommendation model. In a specific application scenario, the inquiry feature information can be input into the inquiry information recommendation model, the model output is used as the information to be inquired, and the information to be inquired is pushed to the inquiring user so that the user can continue to answer the psychological question. After the inquiring user replies to the information to be inquired, the information to be inquired is used as the new inquired information, and the answer of the inquiring user to the question is used as the new answered information, and the new inquired information and the new answered information are added to the inquiry dialogue information, and the process returns to step 102, and the inquiry feature information of the inquiring user is regenerated based on the initial input information and the inquiry dialogue information, and the new inquiry feature information is input into the psychological scale recommendation model, and it is determined whether there is a psychological scale with a matching degree greater than the preset matching degree threshold according to the model output, and if there is, it is recommended to the user, and if not, the above process is repeated until the psychological scale can be recommended to the inquiring user.

[0084] In an embodiment of the present application, optionally, the query information recommendation model includes an input representation layer, a network layer and an output layer, wherein the input representation layer includes a name embedding layer, a role embedding layer and a position embedding layer, and the input representation layer is used to accumulate the respective feature vectors of the name embedding layer, the role embedding layer and the position embedding layer according to the position of each component in the vector; the psychological scale recommendation model includes the input representation layer, the network layer, a fully connected layer and a classification layer.

[0085] In this embodiment, the query information recommendation model and the psychological scale recommendation model have the same model structure. The query information recommendation model includes a name embedding layer, a role embedding layer, a location embedding layer, a network layer, and an output layer. Figure 2 As shown, the name query feature vector [C1, C2, K1, K2, D1, P1, D2, P2, P3] is input into the name embedding layer, the role query feature vector [R1, R1, R2, R2, R3, R4, R3, R4, R4] is input into the role embedding layer, and the location query feature vector [R1, R1, R2, R2, R3, R4, R3, R4, R4] is input into the location embedding layer. The three embedding layers are added according to the corresponding component positions to obtain the input representation layer, and the output layer [D3] is obtained through the network layer. The model structure of the psychological scale recommendation model is the same as that of the inquiry information recommendation model, including the input representation layer, the network layer, the fully connected layer, and the classification layer.

[0086] The embodiment of the present application also provides a model training method for a query information recommendation model and a psychological scale recommendation model. The query information recommendation model and the psychological scale recommendation model have the same network layer parameters, and the training steps of the network layer parameters include:

[0087] Step 201, training the inquiry information recommendation model based on the sample initial input information of the sample inquirer and the sample inquiry dialogue information, and determining the first network layer parameters;

[0088] Step 202, using the first network layer parameters as the network layer parameters of the psychological scale recommendation model, training the psychological scale recommendation model according to the sample initial input information of the sample inquirer, the sample inquiry dialogue information and the sample psychological scale category, and determining the second network layer parameters of the network layer;

[0089] Step 203: Use the second network layer parameters as the network layer parameters of the query information recommendation model, and cyclically train the query information recommendation model and the psychological scale recommendation model until a training end condition is met.

[0090] In this embodiment, a training sample is obtained, and the training sample includes the sample initial entry information of the sample inquirer, the sample inquiry dialogue information, and the sample psychological scale category. For the training of the inquiry information recommendation model, first, the input sample and the corresponding output sample are determined. In addition to the sample initial entry information, the input sample can also intercept part of the dialogue in the sample inquiry dialogue information as the training input, and the inquiry information after this part of the dialogue is used as the training output. For example, if the sample inquiry dialogue information includes 10 inquiry dialogues, the first 5 inquiry dialogues and the sample initial entry information can be intercepted as input samples, and the sixth inquiry question can be used as the output sample. Then, after constructing the input sample and the output sample in the above manner, supervised training is performed on the inquiry information recommendation model to determine the parameters of the network layer, that is, the first network layer parameters. For the training of the psychological scale recommendation model, the network layer of the inquiry information recommendation model can be connected to the fully connected layer and the softmax classification layer for training, and the model is trained using the sample initial entry information, the sample inquiry dialogue information, and the sample psychological scale category to determine the new network layer parameters, that is, the second network layer parameters. Afterwards, the second network layer parameters are used as the network layer parameters of the inquiry information recommendation model, and the above training process for the inquiry information recommendation model and the psychological scale recommendation model is repeated. The two models are alternately trained to determine the final network layer parameters, and the training of the inquiry information recommendation model and the psychological scale recommendation model is completed.

[0091] By applying the technical solution of this embodiment, benefit one: applying the model proposed in this embodiment to recommend intelligent psychological scales replaces part of the work of psychological counselors, greatly saves the manpower of psychological counselors, and saves the time of psychological counseling. Benefit two: The intelligent psychological scale recommendation model proposed in this embodiment has technical advantages. It is based on the existing inquiry information recommendation model technology, and targets specific fields and application scenarios. In this embodiment, i.e., in the online consultation scenario in the field of psychological counseling, the intelligent psychological scale recommendation model structure and the method of training the model are designed through the improvement of the model structure. Benefit three: The capabilities of the model proposed in this embodiment make it beneficial in online consultation scenarios. On the one hand, the model can respond to the inquirer in the form of a dialogue, thereby improving the user experience; on the other hand, the model can conduct in-depth analysis of the information obtained through the dialogue, thereby recommending a more suitable and accurate psychological scale.

[0092] Further, as Figure 1 The specific implementation of the method, the embodiment of the present application provides a psychological scale recommendation device based on artificial intelligence, such as Figure 3 As shown, the device comprises:

[0093] An information acquisition module, used to acquire the initial input information of the inquiring user and the inquiry dialogue information, wherein the inquiry dialogue information includes the inquired information and the answered information;

[0094] A feature determination module, used to determine the inquiry feature information of the inquiring user based on the initial input information and the inquiry dialogue information;

[0095] The recommendation module is used to input the inquiry feature information into the psychological scale recommendation model, and recommend the psychological scale to the inquiring user based on the obtained psychological scale recommendation data.

[0096] Optionally, the initial input information includes user characteristics and query keywords, the user characteristics include the age and gender of the inquiring user, and the query keywords are obtained by matching the query description information of the inquiring user with preset keywords;

[0097] The feature determination module is also used to: generate a name query feature vector, a role query feature vector and a location query feature vector in sequence according to the initial input information and the query dialogue information; accumulate the name query feature vector, the role query feature vector and the location query feature vector according to the position of each component in the vector to determine the query feature information.

[0098] Optionally, the psychological scale recommendation model is used to predict the matching degree between the inquiring user and each psychological scale; and the recommendation module is further used to:

[0099] Determine whether there is any psychological scale in the psychological scale recommendation data whose matching degree with the inquiring user is higher than a preset matching degree threshold;

[0100] If so, the psychological scale that is higher than the preset matching degree threshold is used as the target psychological scale, and the target psychological scale is recommended to the inquiring user.

[0101] Optionally, the recommendation module is further used to:

[0102] After determining whether there is any psychological scale in the psychological scale recommendation data and the inquiring user has a matching degree higher than a preset matching degree threshold, if not, inputting the inquiry feature information into the inquiry information recommendation model to obtain the information to be inquired, outputting the information to be inquired to the inquiring user, and receiving the inquiring user's answer information to the information to be inquired;

[0103] The information to be inquired and the corresponding answer information are added to the inquiry dialogue information as new inquired information and new answered information, and the inquiry feature information of the inquiring user is returned to be re-determined.

[0104] Optionally, the components in the name query feature vector are, in sequence, a user feature component, at least one query keyword component, and at least one group of query dialogue information components, wherein the user feature component includes an age value component and a gender value component, and the query dialogue information component includes a asked information component and an answered information component;

[0105] Each component in the role inquiry feature vector is a feature component symbol corresponding to each component in the name inquiry feature vector, and the role inquiry feature vector includes a user feature symbol, an inquired information symbol, and an answered information symbol;

[0106] Each component in the location query feature vector represents the generation order of each component in the name query feature vector.

[0107] Optionally, the query information recommendation model includes an input representation layer, a network layer and an output layer, wherein the input representation layer includes a name embedding layer, a role embedding layer and a position embedding layer, and the input representation layer is used to accumulate the feature vectors of the name embedding layer, the role embedding layer and the position embedding layer according to the position of each component in the vector;

[0108] The psychological scale recommendation model includes the input representation layer, the network layer, the fully connected layer and the classification layer.

[0109] Optionally, the query information recommendation model and the psychological scale recommendation model have the same network layer parameters, and the device further includes: a model training module, which is used to:

[0110] According to the sample initial input information of the sample inquirer and the sample inquiry dialogue information, the inquiry information recommendation model is trained to determine the first network layer parameters;

[0111] Using the first network layer parameters as the network layer parameters of the psychological scale recommendation model, training the psychological scale recommendation model according to the sample initial input information of the sample inquirer, the sample inquiry dialogue information and the sample psychological scale category, and determining the second network layer parameters of the network layer;

[0112] The second network layer parameters are used as network layer parameters of the query information recommendation model, and the query information recommendation model and the psychological scale recommendation model are cyclically trained until a training end condition is met.

[0113] It should be noted that for other corresponding descriptions of the functional units involved in the artificial intelligence-based psychological scale recommendation device provided in the embodiment of the present application, reference can be made to Figure 1 to Figure 2 The corresponding description in the method will not be repeated here.

[0114] Based on the above Figure 1 to Figure 2The method shown in the embodiment of the present application is accordingly provided with a storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned Figure 1 to Figure 2 The artificial intelligence-based psychological scale recommendation method shown.

[0115] Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each implementation scenario of the present application.

[0116] Based on the above Figure 1 to Figure 2 The method shown, and Figure 3 In order to achieve the above-mentioned purpose, the embodiment of the present application further provides a computer device, which can be a personal computer, a server, a network device, etc. The computer device includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to achieve the above-mentioned Figure 1 to Figure 2 The artificial intelligence-based psychological scale recommendation method shown.

[0117] Optionally, the computer device may further include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, etc. The user interface may include a display, an input unit such as a keyboard, etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (such as a Bluetooth interface, a WI-FI interface), etc.

[0118] Those skilled in the art will appreciate that the computer device structure provided in this embodiment does not limit the computer device, and may include more or fewer components, or a combination of certain components, or different component arrangements.

[0119] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages and saves the hardware and software resources of the computer device, and supports the operation of information processing programs and other software and / or programs. The network communication module is used to realize communication between the components inside the storage medium, and communication with other hardware and software in the physical device.

[0120] Through the description of the above implementation methods, technical personnel in this field can clearly understand that the present application can be implemented by means of software plus necessary general hardware platforms, and can also be implemented by hardware to generate the inquiry feature information of the inquiring user based on the initial input information and inquiry dialogue information of the inquiring user. By pre-constructing a psychological scale recommendation model, the model is used to automatically recommend psychological scales to users according to the inquiry feature information, thereby solving the problem of inaccurate selection of psychological scales by users themselves, and solving the problems of low efficiency and high labor costs caused by psychologists recommending psychological scales, thereby improving the accuracy of psychological scale recommendations and reducing labor costs.

[0121] Those skilled in the art will appreciate that the accompanying drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the accompanying drawings are not necessarily necessary for implementing the present application. Those skilled in the art will appreciate that the modules in the devices in the implementation scenario can be distributed in the devices of the implementation scenario according to the description of the implementation scenario, or can be changed accordingly and located in one or more devices different from the present implementation scenario. The modules of the above-mentioned implementation scenario can be combined into one module, or can be further split into multiple submodules.

[0122] The above serial numbers of this application are only for description and do not represent the advantages and disadvantages of the implementation scenarios. The above disclosure is only a few specific implementation scenarios of this application, but this application is not limited to them, and any changes that can be thought of by technicians in this field should fall within the scope of protection of this application.

Claims

1. A method for recommending psychological scales based on artificial intelligence, characterized in that, the method includes: Obtaining the initial input information of the inquiring user and the inquiry dialogue information, where the inquiry dialogue information includes the information already asked and the information already answered; Determining the inquiry characteristic information of the inquiring user based on the initial input information and the inquiry dialogue information; Inputting the inquiry characteristic information into a psychological scale recommendation model, and based on the obtained psychological scale recommendation data, recommending a psychological scale to the inquiring user, where the psychological scale recommendation model is used to predict the matching degree between the inquiring user and each psychological scale; The recommending a psychological scale to the inquiring user based on the obtained psychological scale recommendation data specifically includes: Judging whether there is any psychological scale in the psychological scale recommendation data whose matching degree with the inquiring user is higher than a preset matching degree threshold; If there is, taking the psychological scale with a matching degree higher than the preset matching degree threshold as the target psychological scale, and recommending the target psychological scale to the inquiring user; If not, inputting the inquiry characteristic information into an inquiry information recommendation model to obtain the information to be asked, outputting the information to be asked to the inquiring user, and receiving the answering information of the inquiring user to the information to be asked; Taking the information to be asked and its corresponding answering information as the new information already asked and the new information already answered, adding them to the inquiry dialogue information, and returning to re-determine the inquiry characteristic information of the inquiring user; wherein, the inquiry information recommendation model includes an input representation layer, a network layer, and an output layer, and the psychological scale recommendation model includes the input representation layer, the network layer, a fully connected layer, and a classification layer; The network layer parameters of the inquiry information recommendation model and the psychological scale recommendation model are the same, and the training steps of the network layer parameters include: Training the inquiry information recommendation model according to the sample initial input information and sample inquiry dialogue information of the sample inquirer to determine the first network layer parameters; Taking the first network layer parameters as the network layer parameters of the psychological scale recommendation model, and training the psychological scale recommendation model according to the sample initial input information, sample inquiry dialogue information, and sample psychological scale categories of the sample inquirer to determine the second network layer parameters of the network layer; Taking the second network layer parameters as the network layer parameters of the inquiry information recommendation model, and cyclically training the inquiry information recommendation model and the psychological scale recommendation model until the training end condition is met.

2. The method according to claim 1, characterized in that, the initial input information includes user characteristics and inquiry keywords, the user characteristics include the age and gender of the inquiring user, and the inquiry keywords are obtained by performing preset keyword matching on the inquiry description information of the inquiring user; The determining the inquiry characteristic information of the inquiring user based on the initial input information and the inquiry dialogue information specifically includes: Generate a name query feature vector, a role query feature vector, and a location query feature vector in sequence according to the initial input information and the query dialogue information; The name query feature vector, the role query feature vector, and the position query feature vector are accumulated according to the position of each component in the vector to determine the query feature information.

3. The method according to claim 2, It is characterized in that The components in the name query feature vector are, in sequence, a user feature component, at least one query keyword component, and at least one group of query dialogue information components, wherein the user feature component includes an age value component and a gender value component, and the query dialogue information component includes a asked information component and an answered information component; Each component in the role inquiry feature vector is a feature component symbol corresponding to each component in the name inquiry feature vector, and the role inquiry feature vector includes a user feature symbol, an inquired information symbol, and an answered information symbol; Each component in the location query feature vector represents the generation order of each component in the name query feature vector.

4. The method according to claim 3, It is characterized in that The input representation layer includes a name embedding layer, a role embedding layer and a position embedding layer, and the input representation layer is used to accumulate the feature vectors of the name embedding layer, the role embedding layer and the position embedding layer according to the position of each component in the vector.

5. A psychological scale recommendation device based on artificial intelligence, applied to the psychological scale recommendation method based on artificial intelligence as claimed in any one of claims 1 to 4, It is characterized in that The device comprises: An information acquisition module, used to acquire the initial input information of the inquiring user and the inquiry dialogue information, wherein the inquiry dialogue information includes the inquired information and the answered information; A feature determination module, used to determine the inquiry feature information of the inquiring user based on the initial input information and the inquiry dialogue information; A recommendation module, used for inputting the inquiry feature information into a psychological scale recommendation model, and recommending a psychological scale to the inquiring user based on the obtained psychological scale recommendation data, wherein the psychological scale recommendation model is used to predict the matching degree between the inquiring user and each psychological scale; If there is no psychological scale with a matching degree greater than the preset matching degree threshold, continue to ask questions to the inquiring user based on the inquiry information recommendation model; The query information recommendation model includes an input representation layer, a network layer, and an output layer, and the psychological scale recommendation model includes the input representation layer, the network layer, a fully connected layer, and a classification layer; The query information recommendation model and the psychological scale recommendation model have the same network layer parameters, and the device further includes: a model training module for: According to the sample initial input information of the sample inquirer and the sample inquiry dialogue information, the inquiry information recommendation model is trained to determine the first network layer parameters; Using the first network layer parameters as the network layer parameters of the psychological scale recommendation model, training the psychological scale recommendation model according to the sample initial input information of the sample inquirer, the sample inquiry dialogue information and the sample psychological scale category, and determining the second network layer parameters of the network layer; The second network layer parameters are used as network layer parameters of the query information recommendation model, and the query information recommendation model and the psychological scale recommendation model are cyclically trained until a training end condition is met.

6. A storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the method for recommending a psychological scale based on artificial intelligence as described in any one of claims 1 to 4 is implemented.

7. A computer device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, It is characterized in that When the processor executes the computer program, the method for recommending a psychological scale based on artificial intelligence as described in any one of claims 1 to 4 is implemented.

Images