Online high-precision psychological evaluation method and system based on big data

Abstract

Embodiments of the present application provide an online high-precision psychological evaluation method and system based on big data, relating to the technical field of psychological evaluation. The online high-precision psychological evaluation method based on big data comprises the following steps: obtaining user information, and generating a type label according to the user information; matching historical evaluation data and evaluation questions of the corresponding type according to the type label, and generating an initial question sequence; the user answers online according to the initial question sequence, and collects feature data in the user's answering process; the current psychological state of the user is obtained based on the feature data, and the initial question sequence is updated according to the user's psychological state to generate a dynamic question sequence; a user psychological evaluation report is generated according to the user's answering result. The problems of large human resource consumption, low preliminary screening efficiency and inaccurate evaluation results in the prior art are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of psychological assessment, and particularly relates to an online high-precision psychological assessment method and system based on big data. BACKGROUND

[0002] In the recruitment process of modern enterprises, evaluating the psychological quality of job applicants has become an important link to ensure the selection of suitable talents. Psychological assessment can comprehensively understand the personality characteristics, stress tolerance, team cooperation ability and other aspects of job applicants, which helps enterprises to screen candidates who meet the job requirements in the recruitment stage, and improves the effectiveness and stability of human resource allocation. Through scientific psychological assessment, enterprises can not only reduce employee turnover rate, but also improve overall work efficiency and organizational cohesion, thereby occupying an advantage in fierce market competition.

[0003] At present, enterprises mainly rely on two methods when conducting psychological assessment: one is through human interviews, and the other is to use fixed test questions for evaluation. Human interviews often require a large amount of human resources, and the subjective judgment of the interviewer may affect the fairness and consistency of the assessment, making it difficult to achieve large-scale preliminary screening. On the other hand, using fixed test questions for psychological assessment has improved the standardization of evaluation to some extent, but since the content of the questions is fixed, job applicants can prepare in advance or simulate practice to cope with the test, which leads to doubts about the accuracy and effectiveness of the test results. SUMMARY

[0004] According to an embodiment of the present application, an online high-precision psychological assessment method and system based on big data are provided to solve the problems raised in the background.

[0005] In a first aspect of the present application, an online high-precision psychological assessment method based on big data is provided.

[0006] The online high-precision psychological assessment method based on big data comprises the following steps:

[0007] S1. Obtain user information and generate a type label according to the user information;

[0008] S2. Obtain cloud data, wherein the cloud data includes historical assessment data and test questions; the historical assessment data and the test questions are divided into at least two types according to the type;

[0009] S3. Match the historical assessment data and the test questions of the corresponding type according to the type label, and generate an initial question sequence;

[0010] S4. The user answers the questions online according to the initial question sequence, and collects feature data in the user's answering process;

[0011] S5. Obtain the current psychological state of the user based on the feature data, and update the initial question sequence according to the psychological state of the user to generate a dynamic question sequence;

[0012] S6. Generate a psychological evaluation report of the user according to the user's answering results.

[0013] Preferably, the step S1 comprises:

[0014] S1-1. Obtain user information, including name, age, gender, occupation and educational background;

[0015] S1-2. Feature encoding is performed on the user information, wherein category variables (such as gender, occupation and educational background) are processed by One-Hot Encoding, and continuous variables (such as age) are standardized;

[0016] S1-3. Weight distribution is performed on the user information after feature encoding, and a type matching score is calculated; the type matching score formula is:

[0017]

[0018] wherein, is the matching score of type T i , F j is the jth encoded user feature, W j is the corresponding weight coefficient, n is the total number of features, and T i is the ith type label;

[0019] S1-4. By comparing the matching scores of each type, the type T k with the highest score is selected as the type label of the user, that is:

[0020] wherein, L is the final user type label, and argmax is the value of i that makes the expression maximum.

[0021] Preferably, the historical evaluation data contains evaluation records of past users, covering user type labels, answer content and emotional response data.

[0022] Preferably, the step S3 comprises:

[0023] S3-1. According to the type label, all evaluation records matching the type label are filtered out from the historical evaluation data stored in the cloud, and the corresponding evaluation questions are extracted;

[0024] S3-2. The filtered evaluation questions are classified according to different psychological dimensions, and each psychological dimension corresponds to an independent question group;

[0025] S3-3. Select a number of questions from each psychological dimension question group in proportion, ensuring that each psychological dimension is represented in the initial question sequence to comprehensively cover the user's psychological state;

[0026] S3-4. Arrange the selected questions in order of importance of each psychological dimension to form an initial question sequence that is logically coherent and structurally reasonable.

[0027] Preferably, the step S4 further comprises:

[0028] S4-1. Classify the feature data;

[0029] S4-2. Establish the association relationship between different categories of features, match and combine the associated feature data through the pre-set association rules to form a composite feature set, so as to reveal the comprehensive behavior pattern of the user in the answering process;

[0030] S4-3. Layered integration of the associated composite feature set according to the psychological dimensions (such as attention, emotional stability, cognitive load), each psychological dimension corresponds to a specific feature subset, and each level of feature is integrated and summarized to generate a comprehensive feature representation of each psychological dimension;

[0031] S4-4. Time series processing of the comprehensive feature representation of each psychological dimension according to the answering order to construct the feature evolution trajectory of the user in the answering process;

[0032] S4-5. Based on the feature evolution trajectory, applying the pre-set psychological state mapping rules to convert the time series comprehensive feature representation into specific psychological state indicators, including but not limited to attention level, emotional stability and cognitive load, etc. to fully reflect the psychological state changes of the user in the answering process.

[0033] Preferably, the step S5 comprises:

[0034] S5-1. Based on the psychological dimension comprehensive feature representation generated in step S4, calculate the current psychological state score P of the user t , wherein t represents the current answering stage;

[0035]

[0036] , wherein m is the total number of psychological dimensions, w i is the weight coefficient of the i-th psychological dimension, and , F i is the feature summary value of the i-th psychological dimension, C i is the feature interaction index of the i-th psychological dimension, D i is the standardization factor of the i-th psychological dimension, a non-linear adjustment factor for the i-th psychological dimension;

[0037] S5-2. According to the current psychological state score of the user, the current psychological state of the user is classified into a first state, a second state and a third state; when the psychological state score is greater than or equal to a first threshold value, the psychological state of the user is the first state; when the psychological state score is greater than or equal to a second threshold value and less than the first threshold value, the psychological state of the user is the second state; and when the psychological state score is less than the second threshold value, the psychological state of the user is the third state;

[0038] S5-3. According to the current psychological state of the user, the questions matching the current psychological state label are filtered from the cloud data, the psychological dimension label of the question is extracted, the questions consistent or highly related to the psychological state label are selected to join the dynamic question sequence, and it is verified whether the selected questions cover all psychological dimensions; if the number of questions of a certain psychological dimension is insufficient, it is marked as a potential missing dimension;

[0039] S5-4. When the number of matching questions is insufficient or the questions of a certain psychological dimension are not fully covered, a new question generation process is triggered, and new questions are generated by calling the question templates corresponding to the missing psychological dimensions; when generating, the label of the current psychological state and the historical answering record of the user are combined to ensure that the content of the new questions is highly related to the psychological state and not repeated; the generated new questions are supplemented to the dynamic question sequence;

[0040] S5-5. The matching questions and the generated new questions are combined, the order of the questions is adjusted to ensure the progression and logical coherence of the psychological dimensions, and it is verified whether the length and the coverage of the psychological dimensions of the sequence meet the evaluation requirements, and finally a dynamic question sequence is generated.

[0041] Preferably, the order of the questions is adjusted by preferentially arranging the questions of the psychological dimensions corresponding to the current psychological state, and inserting the questions of the low-related psychological dimensions between the high-related questions before and after, to maximize the change of psychological load; after each question is inserted, the difference in psychological dimension with the previous question is detected to ensure that the psychological dimensions of every two adjacent questions are alternated minimally, and the questions of the psychological dimensions not touched in the answering record of the user are inserted at the end of the sequence to obtain the dynamic question sequence. The dimension coverage is gradually completed; at the end of the sorting, the positions of the high psychological load questions are adjusted and dispersed to the discontinuous nodes in the dynamic sequence to avoid excessive concentration of psychological load and improve the coherence and adaptability of the evaluation.

[0042] In a second aspect of the present application, an online high-precision psychological evaluation system based on big data is provided.

[0043] The system is used to implement the online high-precision psychological evaluation method described above, comprising:

[0044] a user terminal configured to:

[0045] acquire user information and transmit the user information to a server side;

[0046] receive an initial question sequence generated by the server side and present it to a user;

[0047] collect feature data of the user in the process of answering questions and evaluate the psychological state of the user based on the feature data;

[0048] transmit the psychological state of the user to the server side;

[0049] a server side configured to:

[0050] store historical evaluation data and evaluation questions;

[0051] generate an initial question sequence based on the user information and transmit the initial question sequence to a user side;

[0052] generate a dynamic question sequence based on the psychological state of the user and transmit the dynamic question sequence to the user side.

[0053] Preferably, the user side comprises:

[0054] a display screen for displaying evaluation questions;

[0055] a camera for shooting facial features of the user, acquiring feature data, and being connected to a first data processing unit through an internal data bus;

[0056] an information input module for receiving information input by the user and transmitting it to the server side through a first network communication module;

[0057] a first network communication module for data communication with the server side, the first network communication module being connected to the server side through a network;

[0058] a first data processing unit configured to generate the psychological state of the user according to feature data in the process of answering questions by the user and transmit the psychological state of the user to the server side through the first network communication module.

[0059] the server side comprises:

[0060] a database for storing historical evaluation data and evaluation questions and being connected to a second data processing unit through an internal data bus;

[0061] a second network communication module for data communication with the user side, the second network communication module being connected to the user side through a network;

[0062] a second data processing unit comprising:

[0063] An initial question sequence generation module is configured to filter test questions in a database based on the user information and generate an initial question sequence;

[0064] A dynamic question sequence generation module is configured to generate a dynamic question sequence based on the user's psychological state.

[0065] Preferably, the dynamic question sequence generation module comprises:

[0066] A matching sub-module is configured to filter questions matching the psychological state label from the database according to the user's current psychological state, and add the selected questions to the dynamic question sequence;

[0067] A coverage verification sub-module is configured to verify whether the selected questions cover all psychological dimensions, and if the number of questions for a certain psychological dimension is found to be insufficient, mark the psychological dimension as a potential missing dimension;

[0068] A new question generation sub-module is configured to generate new questions from a question template corresponding to the missing psychological dimension when the number of matching questions is insufficient or a certain psychological dimension is not sufficiently covered, and supplement the new questions to the dynamic question sequence;

[0069] A sorting sub-module is configured to sort the questions in the dynamic question sequence to ensure the gradualness and logical coherence of the psychological dimensions, and verify whether the length and psychological dimension coverage of the sequence meet the test requirements, and finally generate the dynamic question sequence.

[0070] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0071] The online high-precision psychological evaluation method based on big data provided by the application realizes the rapid preliminary screening of large-scale job applicants through automatic user information acquisition and type label generation, significantly reducing the resource investment of enterprises in manpower interviews. Secondly, by using the classified historical evaluation data stored in the cloud and the dynamic question generation mechanism, the personalization and flexibility of the evaluation process are ensured, the question sequence can be dynamically adjusted according to the real-time feature data of the job applicant during the answering process, and the problem that the fixed questions are easily coped with and lead to inaccurate evaluation results is avoided. In addition, the multi-dimensional feature data of the user is deeply analyzed by using big data analysis and machine learning algorithm, which greatly enhances the accuracy and reliability of the evaluation result, ensures the scientificity and practicability of the psychological evaluation report. The interactivity and personalized experience of the online evaluation platform not only improve the participation and evaluation experience of the job applicants, but also provide intuitive and objective recruitment decision basis for enterprises through the visual evaluation report. In summary, the application solves the problems of large manpower consumption, low preliminary screening efficiency and inaccurate evaluation results in the prior art by using an efficient, accurate and dynamic psychological evaluation method, and provides strong technical support for the optimization of enterprise recruitment process and the scientific management of human resources.

[0072] It should be understood that the content described in the summary section is not intended to limit the key or important features of the embodiments of the application, nor to limit the scope of the application. Other features of the application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0073] The above and other features, advantages, and aspects of embodiments of the application will become more apparent by referring to the following detailed description in conjunction with the accompanying drawings, in which like reference numerals indicate like elements, wherein:

[0074] Figure 1 A flowchart of the online high-precision psychological evaluation method based on big data according to an embodiment of the application is shown;

[0075] Figure 2 A flowchart of step S1 of the online high-precision psychological evaluation method based on big data according to an embodiment of the application is shown;

[0076] Figure 3 A flowchart of step S3 of the online high-precision psychological evaluation method based on big data according to an embodiment of the application is shown;

[0077] Figure 4 A flowchart of step S4 of the online high-precision psychological evaluation method based on big data according to an embodiment of the application is shown;

[0078] Figure 5A flow chart of step S5 of the online high-precision psychological evaluation method based on big data according to the embodiment of the present application is shown;

[0079] Figure 6 A block diagram of the online high-precision psychological evaluation system based on big data according to the embodiment of the present application is shown;

[0080] Figure 7 A structural schematic diagram of the user end of the online high-precision psychological evaluation system based on big data according to the embodiment of the present application is shown. DETAILED DESCRIPTION

[0081] To make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0082] In addition, the term "and / or" herein is only used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which can represent the three cases of existence of A alone, existence of A and B together, and existence of B alone. In addition, the character " / " herein generally represents an "or" relationship between the front and rear associated objects.

[0083] As shown in Figure 1 In the first aspect of the present application, an online high-precision psychological evaluation method based on big data is provided.

[0084] The online high-precision psychological evaluation method based on big data comprises the following steps:

[0085] S1. Obtain user information and generate a type label according to the user information; first, obtain user information, take the user information as input data, analyze and process the user information, and generate a corresponding type label for subsequent matching of appropriate evaluation data and evaluation questions. The generation of the type label can adopt a classification algorithm such as a decision tree, a support vector machine (SVM), or a neural network, to ensure the accuracy and effectiveness of the label, thereby providing a personalized basis for psychological evaluation.

[0086] S2. Obtain cloud data, including historical assessment data and assessment questions; the historical assessment data and the assessment questions are classified into at least two categories according to types; the historical assessment data includes the assessment results, answer records and related characteristic data of previous users, which are classified and stored according to different type labels. For example, the data can be classified into multiple categories according to the occupation type, industry background or psychological characteristics of the user. The assessment questions are also classified according to types to ensure that each type of question can effectively evaluate the psychological state of a specific type of user. The storage and management of cloud data use a distributed database system to ensure high availability and scalability of data.

[0087] S3. Match the corresponding type of historical assessment data and assessment questions according to the type label, and generate an initial question sequence; after obtaining the type label of the user, the system will match the corresponding type of historical assessment data and assessment questions from the cloud data according to these labels. The matching process involves screening data and questions highly related to the user's type label to ensure the relevance and effectiveness of the assessment. By sorting and combining the matched assessment questions, the system generates an initial question sequence that aims to comprehensively cover various aspects of the user's psychological characteristics. The generation of the initial question sequence can take into account factors such as question difficulty, question type diversity and assessment goals to optimize the assessment process and user experience.

[0088] S4. The user answers the questions online according to the initial question sequence and collects characteristic data during the user's answering process, wherein the characteristic data includes but is not limited to:

[0089] Answering time: records the answering time of each question of the user, reflecting the user's thinking speed and decision-making efficiency.

[0090] Answering accuracy: statistics the answering accuracy of the user, evaluating the user's knowledge mastery and understanding ability.

[0091] Facial expression changes: monitor the user's emotional fluctuations during the answering process through the camera, such as tension, relaxation, etc.

[0092] Behavioral characteristics: record the user's mouse click frequency, keyboard input speed, etc., reflecting the user's concentration and operation habits.

[0093] S5. Based on the feature data, the current psychological state of the user is obtained, and according to the user's psychological state, the initial question sequence is updated to generate a dynamic question sequence; in the process of user answering, based on the feature data of the user when answering, big data analysis technology and machine learning algorithm such as clustering analysis, regression analysis or deep learning model are used to evaluate the current psychological state of the user. Based on the real-time evaluation result, the selection and order of subsequent questions are adjusted in real time to generate a dynamic question sequence. This dynamic adjustment mechanism enables the evaluation process to push personalized questions according to the real-time performance and psychological state of the user, thereby deeply exploring the psychological characteristics of the user. The generation of dynamic question sequence not only considers the relevance and difficulty of the questions, but also optimizes the evaluation path according to the real-time feedback of the user, ensuring the comprehensiveness and depth of the evaluation, and improving the accuracy and reliability of the evaluation result.

[0094] S6. According to the user's answer results, a user psychological evaluation report is generated. After completing all the evaluation questions, the system will comprehensively analyze the user's answer results and feature data to generate a detailed psychological evaluation report. The report includes the user's scores in various psychological dimensions, personality characteristic analysis, stress tolerance evaluation, team cooperation ability analysis and other contents. The evaluation report adopts a combination of visual charts and textual descriptions to intuitively display the user's psychological state and potential strengths and weaknesses. The report can also provide personalized suggestions and improvement measures to help users better understand their psychological characteristics and provide scientific recruitment decision basis for enterprises. The whole evaluation report generation process relies on big data analysis and intelligent algorithm to ensure the objectivity, accuracy and practicality of the report content.

[0095] The online high-precision psychological evaluation method based on big data provided by the application significantly improves the efficiency and accuracy of the recruitment process. First, through automatic user information acquisition and type label generation, rapid preliminary screening of large-scale job seekers is realized, significantly reducing the resource investment of enterprises in manpower interviews. Second, by using the classified historical evaluation data stored in the cloud and the dynamic question generation mechanism, the personalization and flexibility of the evaluation process are ensured, and the question sequence can be dynamically adjusted according to the real-time feature data of the job seeker during the answering process, avoiding the problem of inaccurate evaluation results caused by fixed questions being easily dealt with. In addition, big data analysis and machine learning algorithms are used to deeply analyze the multi-dimensional feature data of users, greatly enhancing the accuracy and reliability of the evaluation results, ensuring the scientificity and practicality of the psychological evaluation report. The interactivity and personalized experience of the online evaluation platform not only improve the participation and evaluation experience of the job seekers, but also provide intuitive and objective recruitment decision-making basis for enterprises through visual evaluation reports. In summary, the application solves the problems of high manpower consumption, low preliminary screening efficiency, and inaccurate evaluation results in the prior art through an efficient, accurate, and dynamic psychological evaluation method, providing strong technical support for the optimization of enterprise recruitment processes and the scientific management of human resources.

[0096] As Figure 2 shown in the present embodiment, the step S1 includes:

[0097] S1-1. Obtain user information, including name, age, gender, occupation, and educational background;

[0098] S1-2. Feature encode the user information, where category variables (such as gender, occupation, and educational background) are processed by One-Hot Encoding, and continuous variables (such as age) are standardized. Further, in order to facilitate subsequent algorithm processing and analysis, the collected user information needs to be feature encoded. The specific processing method is as follows:

[0099] Category variable encoding: For category variables (such as gender, occupation, and educational background), the One-Hot Encoding method is used. One-Hot Encoding converts each category variable into multiple binary variables, each variable corresponding to a category. For example, the gender variable "male" and "female" will be encoded as 1,0 and 0,1 respectively.

[0100] Continuous variable standardization: For continuous variables (such as age), standardization is used. Standardization converts continuous variables into a standard normal distribution with a mean of 0 and a standard deviation of 1, to eliminate the influence of different variable dimensions and improve the training effect of the model. The standardization formula is as follows:

[0101]

[0102] wherein, X is the original age data, is the mean of the age data, is the standard deviation of the age data.

[0103] Through feature encoding, the original user information is converted into a numerical feature vector, which is convenient for subsequent algorithm processing and analysis.

[0104] S1-3. Assign weights to the user information after feature encoding, and calculate the type matching score; the type matching score formula is:

[0105]

[0106] wherein, is the matching score of type T i , F j is the jth encoded user feature, W j is the corresponding weight coefficient, n is the total number of features, T i is the ith type label;

[0107] By weighted summation, the contribution of each feature to the type matching is integrated to obtain the matching score of each type. The weight coefficient is assigned according to the influence degree of each feature on the psychological state

[0108] S1-4. After obtaining the matching scores of all types, the system needs to determine the type label that best matches the user features. Compare the matching scores of the user on each type, and select the type Ti with the highest score as the type label of the user, that is:

[0109] wherein, L is the final user type label, and argmax is the value of i that makes the expression maximum.

[0110] In this embodiment, the historical evaluation data includes the evaluation records of past users, covering user type labels, answer content, and emotional response data.

[0111] As Figure 3 shown, in this embodiment, the step S3 includes:

[0112] S3-1. According to the type label, all evaluation records matching the type label are filtered out from the historical evaluation data stored in the cloud, and the corresponding evaluation questions are extracted; using the type label of the user, all evaluation records matching the type label are filtered out from the historical evaluation data. The historical evaluation data includes the evaluation records of past users, each record covering the type label of the user, the answer content and the emotional response data. The type label is used to identify the psychological type of the user, the answer content records the user's answers to each evaluation question, and the emotional response data reflects the user's emotional changes during the evaluation process. According to the type label of the current user, the system finds the historical evaluation records corresponding to the type label, which can be realized by database query statements such as SELECT statements in SQL, filtering out all records that meet the type label matching conditions. From the filtered matching evaluation records, the corresponding evaluation questions are extracted. Since different records may contain different questions, it is necessary to ensure that the extracted questions cover a variety of types in order to facilitate subsequent classification and selection. In the extraction process, it also includes removing duplicate questions and retaining a unique set of questions to optimize the diversity and coverage of the question bank.

[0113] S3-2. The filtered evaluation questions are classified according to different psychological dimensions, and each psychological dimension corresponds to an independent question group; further, a plurality of psychological dimensions are defined in advance, each dimension corresponding to a specific psychological trait or behavior pattern. For example, common psychological dimensions include emotional stability, extroversion, responsibility, openness and agreeableness, etc. According to the content and evaluation goal of each evaluation question, it is assigned to the corresponding psychological dimension group. This process can be realized by keyword matching, topic analysis or expert annotation, etc., to ensure that each question is correctly classified into the corresponding psychological dimension. Each psychological dimension corresponds to an independent question group, ensuring that the questions of each dimension can independently and comprehensively evaluate the user's psychological traits when selecting questions later.

[0114] S3-3. A number of questions are selected from each psychological dimension question group in proportion, ensuring that each psychological dimension has a representative in the initial question sequence to comprehensively cover the user's psychological state; further, according to the importance and evaluation needs of different psychological dimensions, the proportion of each dimension in the initial question sequence is set in advance. For example, some dimensions may require more questions to obtain more detailed evaluation. In each psychological dimension question group, a number of questions are randomly selected according to the set proportion. Random selection helps to increase the diversity of the evaluation and avoid the influence of the order of the questions on the evaluation results. By selecting in proportion, it is ensured that each psychological dimension has representative questions in the initial question sequence, thereby comprehensively covering the user's various psychological dimensions and improving the comprehensiveness and accuracy of the evaluation.

[0115] S3-4. According to the importance order of each psychological dimension, arrange the selected questions to form an initial question sequence that is logically coherent and structurally reasonable. According to the evaluation target and application scenario, set the importance order of each psychological dimension. For example, in some job recruitment, responsibility and team cooperation ability may be more important than openness. According to the importance order of each psychological dimension, arrange the selected questions. Prioritize the arrangement of questions corresponding to dimensions with higher importance to ensure that key psychological characteristics are fully evaluated. In the process of question arrangement, consider the logical relationship between questions and the coherence of the evaluation process to avoid abrupt jumps or repetitions between questions, and improve the user's answering experience. Ensure that the initial question sequence is reasonably distributed in structure to avoid excessive concentration or dispersion of questions on the same psychological dimension, and maintain the balance and systematicness of the evaluation process.

[0116] As shown in Figure 4 , in the present embodiment, the step S4 further comprises:

[0117] S4-1. Classify the feature data; the feature data includes but is not limited to the answering time, emotional data. Reflect the behavior and psychological reaction of the user in the answering process.

[0118] S4-2. Establish the association relationship between different categories of features, match and combine the associated feature data according to the preset association rules to form a composite feature set to reveal the comprehensive behavior pattern of the user in the answering process; by establishing the association relationship between different categories of features, a composite feature set is generated to reveal the comprehensive behavior pattern of the user in the answering process. Emotional fluctuations may be related to the increase of cognitive load (such as the extension of answering time), and according to the preset association rules, the associated feature data is matched and combined. For example: if the user shows emotional fluctuations in multiple consecutive questions, these emotion-related features will be combined into a composite feature "emotional instability". Through the matching and combination of the association rules, multiple composite feature sets are generated. These composite feature sets can more comprehensively reflect the comprehensive behavior pattern of the user in the answering process, surpassing the analysis of single features and providing more abundant information for subsequent psychological state evaluation.

[0119] S4-3. According to the psychological dimensions (such as attention, emotional stability, cognitive load), the associated composite feature set is integrated in layers, each psychological dimension corresponds to a specific feature subset, and each layer of features is integrated and summarized to generate a comprehensive feature representation of each psychological dimension.

[0120] Furthermore, a plurality of psychological dimensions are predefined, each dimension corresponding to a specific psychological trait or behavior pattern. For example:

[0121] Attention: reflects the degree of concentration of the user in the answering process.

[0122] Emotional stability: reflects the emotional fluctuations of the user during the answering process.

[0123] Cognitive load: reflects the mental burden and processing capacity of the user during the answering process.

[0124] According to the definition of each psychological dimension, the associated composite feature set is divided into corresponding feature subsets. For example:

[0125] Attention feature subset: includes the stability of answering time, the consistency of page dwell time, etc.

[0126] Emotional stability feature subset: includes the frequency and amplitude of emotional fluctuations in emotional response data, etc.

[0127] Cognitive load feature subset: includes changes in answering speed, extension of thinking time, etc.

[0128] For each psychological dimension feature subset, the features are summarized and integrated using the integration method, and the comprehensive feature representation of each psychological dimension is generated.

[0129] S4-4. The comprehensive feature representation of each psychological dimension is time-sequenced according to the answering order, and the feature evolution trajectory of the user during the answering process is constructed.

[0130] Further, the comprehensive feature representation of the user on each question is regarded as a time point data, arranged according to the user's answering order to form a time sequence. For each psychological dimension, the system connects its comprehensive feature representation according to the answering order to form a continuous time sequence. For example:

[0131] Attention comprehensive feature representation of the user on the first to nth question.

[0132] Emotional stability comprehensive feature representation of the user on the first to nth question.

[0133] Cognitive load comprehensive feature representation of the user on the first to nth question.

[0134] The comprehensive feature representation of each question is arranged in sequence using a sliding window or cumulative method to ensure the continuity and integrity of the feature evolution trajectory. The constructed feature evolution trajectory is stored as multi-dimensional time series data for subsequent analysis and mapping. For example, the attention, emotional stability and cognitive load trajectory data can be stored as three independent time series, or they can be integrated into a multi-dimensional time series.

[0135] S4-5. Based on the feature evolution trajectory, apply the preset psychological state mapping rules to convert the time-series comprehensive feature representation into specific psychological state indicators, including but not limited to attention level, emotional stability, and cognitive load, to comprehensively reflect the user's psychological state changes during the answering process.

[0136] Further, a set of psychological state mapping rules are preset for converting the comprehensive feature representation into specific psychological state indicators. These rules can be based on psychological theories, expert experience, or data-driven methods. For example:

[0137] Attention level mapping: If the attention comprehensive feature is consistently above a certain threshold at multiple time points, it is mapped to "high attention". If the attention comprehensive feature is below a certain threshold at multiple time points, it is mapped to "low attention".

[0138] Emotional stability mapping: If the emotional stability comprehensive feature has a small fluctuation range throughout the answering process, it is mapped to "emotional stability". If the emotional stability comprehensive feature has a large fluctuation range, it is mapped to "emotional instability".

[0139] Cognitive load mapping: If the cognitive load comprehensive feature is above a certain threshold at most time points, it is mapped to "high cognitive load". If the cognitive load comprehensive feature is below a certain threshold at most time points, it is mapped to "low cognitive load".

[0140] Attention level: such as "high attention", "medium attention", "low attention".

[0141] Emotional stability: such as "emotional stability", "emotional fluctuation".

[0142] Cognitive load: such as "high cognitive load", "medium cognitive load", "low cognitive load".

[0143] Synthesize the psychological state indicators of each psychological dimension to form the overall psychological state changes of the user during the answering process. For example, the user may exhibit different psychological state combinations such as "high attention, emotional stability, low cognitive load" or "low attention, emotional fluctuation, high cognitive load".

[0144] By classifying, associating, integrating, and time-series processing multi-dimensional feature data collected from users' online quiz responses, and applying psychological state mapping rules based on feature evolution trajectories, specific psychological state indicators are generated. This process not only deeply reveals the comprehensive behavioral patterns and psychological state changes of users during the quiz process but also provides rich and accurate data support for subsequent dynamic question generation and the generation of the final psychological assessment report. Through scientific feature processing and psychological state mapping mechanisms, the efficiency, accuracy, and practicality of the psychological assessment method are ensured.

[0145] like Figure 5 As shown, in this embodiment, step S4 has generated comprehensive feature representations of various psychological dimensions (such as attention, emotional stability, cognitive load, etc.). These comprehensive feature representations can fully reflect the user's performance and changes in various psychological dimensions. Step S5 includes:

[0146] S5-1. Based on the comprehensive feature representation of psychological dimensions generated in step S4, calculate the user's current psychological state score P. t , where t represents the current answering stage;

[0147]

[0148] Where m is the total number of psychological dimensions, w i Let be the weight coefficient of the i-th psychological dimension, and F i C represents the feature summary value of the i-th psychological dimension. i D is the feature interaction index for the i-th psychological dimension. i Let i be the standardized factor for the i-th psychological dimension. This is the non-linear adjustment factor for the i-th psychological dimension. Through the above formula, the psychological state score is calculated, ensuring the dynamism and accuracy of the psychological assessment and providing a scientific basis for the dynamic adjustment of the subsequent question sequence.

[0149] S5-2. Based on the user's current psychological state score, classify the user's current psychological state into three states: first state, second state, and third state. When the psychological state score is greater than or equal to the first threshold, the user's psychological state is in the first state. When the psychological state score is greater than or equal to the second threshold and less than the first threshold, the user's psychological state is in the second state. When the psychological state score is less than the second threshold, the user's psychological state is in the third state.

[0150] The system compares the user's psychological state scores, determines the score range in which the psychological state score falls, and marks the user's psychological state as the first state, second state, or third state based on the score range.

[0151] S5-3. According to the current psychological state of the user, filter the questions matching the current psychological state label from the cloud data. By extracting the psychological dimension labels of the questions, select the questions consistent or highly related to the psychological state label and add them to the dynamic question sequence. At the same time, verify whether the selected questions cover all psychological dimensions. If the number of questions for a certain psychological dimension is insufficient, mark it as a potential missing dimension.

[0152] Further, according to the current psychological state classification result of the user (first state, second state or third state), filter the evaluation questions matching the state label from the cloud data. Each evaluation question is pre-labeled with the corresponding psychological dimension label. Extract these labels to ensure that the questions are highly related to the user's psychological state. Select questions consistent or highly related to the current psychological state label to ensure the relevance and effectiveness of the evaluation. During the selection of questions, it is necessary to verify whether the selected questions cover all psychological dimensions. If the number of questions for a certain psychological dimension is insufficient, the system will mark it as a potential missing dimension, ready for supplementation in subsequent steps.

[0153] S5-4. When the number of matching questions is insufficient or the questions for a certain psychological dimension are not fully covered, trigger the new question generation process, generate new questions by calling the question templates corresponding to the missing psychological dimensions; when generating, combine the label of the current psychological state and the user's historical answer record to ensure that the content of the new questions is highly related to the psychological state and not repeated; supplement the generated new questions to the dynamic question sequence.

[0154] Further, in the case where the number of matching questions is insufficient or the questions for a certain psychological dimension are not fully covered, start the new question generation process to supplement the dynamic question sequence. According to the potential missing dimensions marked in step S5-3, determine the psychological dimensions that need to be supplemented, call the question templates corresponding to the missing psychological dimensions, generate new questions through the preset question generation algorithm, and supplement the generated new questions to the dynamic question sequence to ensure that the number and coverage of questions for all psychological dimensions meet the preset standards. At the same time, the generated new questions will be supplemented to the cloud data to realize the update of the cloud data.

[0155] S5-5. Merge the matching questions and the generated new questions, adjust the order of the questions to ensure the gradualness and logical coherence of the psychological dimensions, and verify whether the length and psychological dimension coverage of the sequence meet the evaluation requirements, finally generate the dynamic question sequence.

[0156] The selected matching questions are combined with the generated new questions, and the order of the questions is adjusted to ensure the progression of psychological dimensions and the logical coherence of the assessment process. The matching questions selected in step S5-3 are combined with the new questions generated in step S5-4 to form a complete set of questions. According to the importance of each psychological dimension and the requirements of the assessment process, the progressive order of the questions is set. For example, first assess the basic psychological traits, and then deeply assess the complex psychological state. Ensure the logical coherence between questions in terms of content and assessment goals, avoid abrupt jumps in question order, and improve the user's answering experience.

[0157] Verify whether the total length of the dynamic question sequence meets the assessment requirements, to avoid user fatigue caused by too many questions or impact on assessment accuracy caused by too few questions. Check if the question sequence covers all the preset psychological dimensions to ensure the comprehensiveness of the assessment. Through the above sorting and verification process, the system finally generates a dynamic question sequence that meets the assessment requirements, ensuring the scientificity of the assessment process and the optimization of user experience.

[0158] Based on the real-time psychological state score of the user during the answering process, the initial question sequence is dynamically updated to generate an adaptive dynamic question sequence. This process not only realizes the personalization and precision of the assessment, but also improves the comprehensiveness of the assessment and the user's answering experience through diversity and coverage. The introduction of dynamic adjustment mechanism ensures the real-time and flexibility of the assessment process, providing more rich and accurate data support for the final psychological assessment report generation.

[0159] For example, before the assessment begins, basic information needs to be filled out, including name "Wang", age "28 years old", gender "female", occupation "data analyst", and education background "master". These information is collected by the platform's user information acquisition module and transmitted to the assessment system.

[0160] The user information is feature encoded, the gender "female" is converted to 0, 1 through one-hot encoding; the occupation "data analyst" has multiple possible categories, assuming there are five types of occupations, encoded as 0, 0, 1, 0, 0; the education background "master" assumes there are three levels of education, encoded as 0, 1, 0.

[0161] The age of 28 years old is processed by standardization, assuming that the mean μ of age data is 30 years old and the standard deviation σ is 5 years old, then the standardized age data is:

[0162]

[0163] Through the above encoding, Wang's original user information is transformed into a numerical feature vector: [0, 1, 0, 0, 1, 0, -0.4], where each dimension corresponds to the one-hot encoding of gender, occupation, and education background, and the standardized age. According to the pre-set weight distribution mechanism, each encoded user feature is assigned a corresponding weight coefficient. Assume the weight coefficients are as follows:

[0164] The gender-related feature weight W1 is 0.2;

[0165] The occupation-related feature weight W2 is 0.3;

[0166] The education background-related feature weight W3 is 0.3;

[0167] The age feature weight W4 is 0.2;

[0168] For each type T i , calculate its matching score S i ; assume the system pre-sets three type labels: Type A, Type B, and Type C, each corresponding to different feature weights and matching standards

[0169] The calculation is as follows:

[0170]

[0171] Through weighted summation, the matching scores of each type are obtained: Type A is 0.22, Type B is 0.42, and Type C is 0.22.

[0172] The system compares the matching scores of all types and selects the type with the highest score as the user's type label. For Wang, the matching score of Type B is the highest, 0.42, so the final user type label is Type L = Type B.

[0173] Obtain historical assessment data and assessment questions from the cloud database. These data are divided into at least two categories according to the type, ensuring that the assessment data and questions related to Type B are prioritized.

[0174] According to the type label B, the system selects the assessment records matching Type B from the historical assessment data, extracts the corresponding assessment questions, and classifies them according to different psychological dimensions (such as logical thinking, stress management, and innovation ability). The system selects a certain number of questions from each psychological dimension question group in proportion to ensure that each dimension is represented in the initial question sequence, and finally generates a logically coherent initial question sequence.

[0175] The system displays the initial question through the display screen 11 and captures the facial features of Wang through the camera 12 in real time to collect facial expression changes and other feature data. The information input module 13 receives Wang's answer input and transmits the user information and feature data to the server end 2 through the network communication module 14.

[0176] According to the collected feature data, the psychological state score of Wang in the current answering stage is calculated, assuming that there are three psychological dimensions: attention (weight coefficient 0.3), emotional stability (weight coefficient 0.4), and cognitive load (weight coefficient 0.3).

[0177] The feature summary value of Wang in each dimension is:

[0178] Attention F1=0.8;

[0179] Emotional stability F2=0.6;

[0180] Cognitive load F3=0.7;

[0181] Feature interaction index C i And the standardization factor D i Respectively:

[0182] C1=0.2, D1=1.1

[0183] C2=0.1, D2=0.9

[0184] C3=0.3, D3=1.0;

[0185] The final psychological state score is 1.3. The first threshold is set to 1.5 and the second threshold is set to 1.0, so the current psychological state is the second state.

[0186] From the cloud database, select the questions that match the second state, extract the psychological dimension labels of these questions (such as stress management, logical thinking, etc.), and check if they cover all psychological dimensions. It is found that the number of questions in the "innovation ability" dimension is insufficient, which is marked as a potential missing dimension. Then call the question template corresponding to "innovation ability" to generate new related questions and supplement them to the dynamic question sequence to ensure that all psychological dimensions are covered by enough questions.

[0187] Prioritize the "stress management" questions corresponding to the current psychological state "second state", and insert the "logical thinking" questions with low correlation between the "stress management" questions to maximize the change in psychological load. After each insertion, the system detects the difference in psychological dimension from the previous question to ensure that the psychological dimensions of adjacent questions alternate minimally. The "innovation ability" questions that have not been touched are inserted at the end of the sequence to gradually complete the dimension coverage. At the end of the sorting, the system disperses the high psychological load questions (such as "stress management" questions) to non-continuous nodes in the sequence to avoid excessive concentration of psychological load and improve the continuity and adaptability of the evaluation. Finally, a logically coherent and reasonably distributed psychological load dynamic question sequence is generated and transmitted back to the user end 1. After completing all the evaluation questions, the user's answer results and real-time psychological state scores are analyzed to generate a detailed psychological evaluation report.

[0188] In the present embodiment, the adjustment of the question order is achieved by prioritizing the questions corresponding to the current psychological state, and inserting questions of low-correlation psychological dimensions between high-correlation questions to maximize the change in psychological load. After each insertion, the difference in psychological dimension from the previous question is detected to ensure that the psychological dimensions of every two adjacent questions alternate minimally. Questions of psychological dimensions that have not been touched in the user's answer record are inserted at the end of the sequence to obtain a dynamic question sequence. Gradually complete the dimension coverage; At the end of the sorting, adjust the position of high psychological load questions and disperse them to non-continuous nodes in the dynamic sequence to avoid excessive concentration of psychological load and improve the continuity and adaptability of the evaluation.

[0189] As shown in Figure 6 , in the second aspect of the present application, an online high-precision psychological evaluation system based on big data is provided.

[0190] The system is used to implement the online high-precision psychological evaluation method described above, comprising:

[0191] The user end 1 is configured to:

[0192] Obtain user information and transmit the user information to the server end 2;

[0193] Receive the initial question sequence generated by the server end 2 and present it to the user;

[0194] Collect feature data of the user during the answering process and evaluate the user's psychological state based on the feature data;

[0195] Transmit the user's psychological state to the server end 2;

[0196] The server end 2 is configured to:

[0197] Store historical evaluation data and evaluation questions;

[0198] Generate an initial question sequence based on the user information and transmit the initial question sequence to the user terminal 1;

[0199] Generate a dynamic question sequence based on the user's psychological state and transmit the dynamic question sequence to the user terminal 1.

[0200] In actual use, the user terminal 1 is responsible for collecting the user's basic information and transmitting it to the server terminal 2 through a secure transmission protocol; the server terminal 2 generates an initial question sequence based on the received user information, and dynamically generates an adjusted question sequence based on the user's real-time psychological state during the answering process, to ensure the personalization and accuracy of the evaluation. The server terminal 2 transmits the initial question sequence to the user terminal 1, which receives and displays the initial and dynamic question sequences, guiding the user to complete the online answering. In actual use, the user terminal 1 displays the questions through a graphical user interface and collects the user's answering behavior data, such as answering time, answering order, and interaction behavior, and transmits these feature data to the server terminal 2 in real time; the server terminal 2 evaluates the user's psychological state based on the received feature data, adjusts the question sequence and feeds back to the user terminal 1, realizing dynamic optimization and continuous personalized adjustment of the evaluation process, ensuring the efficiency and scientificity of the entire psychological evaluation process.

[0201] As shown in Figure 7 , in this embodiment, the user terminal 1 comprises:

[0202] A display screen 11 for displaying evaluation questions;

[0203] A camera 12 for shooting the user's facial features, obtaining feature data, and connecting to the first data processing unit 15 through an internal data bus;

[0204] An information input module 13 for receiving user input information and transmitting it to the server terminal 2 through the first network communication module 14;

[0205] A first network communication module 14 for data communication with the server terminal 2, the first network communication module 14 being connected to the server terminal 2 through a network;

[0206] A first data processing unit 15 configured to generate the user's psychological state based on the feature data during the user's answering process and transmit the user's psychological state to the server terminal 2 through the first network communication module 14.

[0207] The server terminal 2 comprises:

[0208] A database 21 for storing historical evaluation data and evaluation questions, and connecting to the second data processing unit 23 through an internal data bus;

[0209] The second network communication module 22 is configured to communicate data with the user terminal 1, and the second network communication module 22 is connected with the user terminal 1 through a network;

[0210] The second data processing unit 23 comprises:

[0211] The initial question sequence generation module 151 is configured to screen the evaluation questions in the database 21 based on the user information and generate an initial question sequence;

[0212] The dynamic question sequence generation module 152 is configured to generate a dynamic question sequence based on the user psychological state.

[0213] In actual use, the user terminal 1 displays the evaluation questions through the display screen 11, the camera 12 collects the facial feature data of the user in real time, the information input module 13 receives the answer input of the user, the first data processing unit 15 evaluates the psychological state of the user according to the collected feature data and transmits the psychological state to the server terminal 2; the server terminal 2 generates an initial question sequence through the initial question sequence generation module 151 by using the historical evaluation data and the evaluation questions in the database 21 and transmits the initial question sequence to the user terminal 1. In the process of answering the questions by the user, the dynamic question sequence generation module 152 adjusts the subsequent evaluation question sequence in real time according to the received psychological state score of the user, ensures the individualization and accuracy of the evaluation, and thus realizes an efficient and scientific psychological evaluation process.

[0214] In the embodiment, the dynamic question sequence generation module 152 comprises a matching sub-module for screening questions matching the psychological state label from the database 21 according to the current psychological state of the user and adding the selected questions to the dynamic question sequence; a coverage verification sub-module for verifying whether the selected questions cover all psychological dimensions, and marking a psychological dimension as a potential missing dimension if the number of questions for the psychological dimension is found to be insufficient; a new question generation sub-module for generating new questions corresponding to the missing psychological dimension by calling the question templates corresponding to the missing psychological dimension when the number of matching questions is insufficient or the questions for a psychological dimension are not sufficiently covered, and supplementing the new questions to the dynamic question sequence; and a sorting sub-module for sorting the questions in the dynamic question sequence to ensure the progression of psychological dimensions and logical coherence, and verifying whether the length and psychological dimension coverage of the sequence meet the evaluation requirements, and finally generating the dynamic question sequence. In actual use, the user end 1 displays the evaluation questions through the display screen 11, the camera 12 collects the facial feature data of the user in real time, the information input module 13 receives the user's answer input, and the first network communication module 14 transmits the user information and feature data to the server end 2 safely; the server end 2 generates an initial question sequence by using the historical evaluation data and evaluation questions in the database 21 through the initial question sequence generation module 151, and transmits the initial question sequence to the user end 1. During the user's answering process, the first data processing unit 15 evaluates the psychological state score of the user according to the collected feature data, and transmits the score to the server end 2 through the first network communication module 14. The server end 2 dynamically adjusts the subsequent evaluation question sequence by screening, verifying, generating and sorting the questions based on the psychological state score through the sub-modules in the dynamic question sequence generation module 152, ensures the individualization and accuracy of the evaluation, and transmits the adjusted dynamic question sequence back to the user end 1. The user end 1 receives and displays the initial and dynamic question sequences, guides the user to complete the online answering, and the whole evaluation process realizes the efficient cooperation between the user end and the server end, ensures the scientificity, real-time performance of the psychological evaluation and the optimization of the user experience.

[0215] The foregoing detailed description does not constitute a limitation of the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A high-precision online psychological assessment method based on big data, characterized in that: Includes the following steps: S1. Obtain user information and generate type tags based on the user information; S2. Obtain cloud data, which includes historical assessment data and assessment questions; the historical assessment data and assessment questions are divided into at least two categories according to type; S3. Match the corresponding historical assessment data and assessment questions according to the type label, and generate an initial question sequence; S4. The user answers questions online according to the initial question sequence, and the feature data of the user during the question-answering process is collected; S5. Based on the feature data, obtain the user's current psychological state, and update the initial question sequence according to the user's psychological state to generate a dynamic question sequence; S6. Generate a user psychological assessment report based on the user's answers; Step S1 includes: S1-1. Obtain user information, including name, age, gender, occupation, and educational background; S1-2. Perform feature encoding on the user information; S1-3. Weights are assigned to the user information after feature encoding, and a type matching score is calculated; the formula for the type matching score is: , in, For type T i The matching score, F j For the j-th encoded user feature, W j Here are the corresponding weight coefficients, where n is the total number of features, and T is the total number of features. i For the i-th type of label; S1-4. By comparing the matching scores of each type, select the type T with the highest score. k As a type label for users, that is: , Where L is the final user type label, and argmax is the value of i that maximizes the expression; Step S3 includes: S3-1. Based on the type tag, filter out all assessment records that match the type tag from the historical assessment data stored in the cloud, and extract the corresponding assessment questions; S3-2. Categorize the selected assessment questions according to different psychological dimensions, with each psychological dimension corresponding to an independent question group; S3-3. Select a number of questions proportionally from the question group of each psychological dimension to ensure that each psychological dimension is represented in the initial question sequence, so as to fully cover the user's psychological state; S3-4. Arrange the selected questions according to the order of importance of each psychological dimension to form a logically coherent and structurally sound initial question sequence.

2. The online high-precision psychological assessment method based on big data as described in claim 1, characterized in that, The historical assessment data includes past user assessment records, covering user type tags, answer content, and emotional response data.

3. The online high-precision psychological assessment method based on big data according to claim 1, characterized in that, Step S4 further includes: S4-1. Classify the feature data; S4-2. Establish the association between different categories of features, and match and combine related feature data through preset association rules to form a composite feature set; S4-3. The composite feature set after association is hierarchically integrated according to the psychological dimension. Each psychological dimension corresponds to a specific feature subset. The features of each level are integrated and summarized to generate a comprehensive feature representation of each psychological dimension. S4-4. The comprehensive feature representations of each psychological dimension are processed in time sequence according to the order of answering questions to construct the feature evolution trajectory of users in the process of answering questions; S4-5. Based on the feature evolution trajectory, apply the preset psychological state mapping rules to transform the time-series comprehensive feature representation into specific psychological state indicators.

4. The online high-precision psychological assessment method based on big data according to claim 3, characterized in that, Step S5 includes: S5-1. Based on the comprehensive feature representation of psychological dimensions generated in step S4, calculate the user's current psychological state score P. t , where t represents the current answering stage; , Where m is the total number of psychological dimensions, w i Let be the weight coefficient of the i-th psychological dimension, and F i C represents the feature summary value of the i-th psychological dimension. i D is the feature interaction index for the i-th psychological dimension. i Let i be the standardized factor for the i-th psychological dimension. For the i-th psychological dimension, there is a non-linear adjustment factor. S5-2. Based on the user's current psychological state score, classify the user's current psychological state into three states: first state, second state, and third state. When the psychological state score is greater than or equal to the first threshold, the user's psychological state is the first state; when the psychological state score is greater than or equal to the second threshold and less than the first threshold, the user's psychological state is the second state; when the psychological state score is less than the second threshold, the user's psychological state is the third state. S5-3. Based on the user's current psychological state, filter questions that match the current psychological state label from cloud data. By extracting the psychological dimension labels of the questions, select questions that are consistent with or highly related to the psychological state label and add them to the dynamic question sequence. At the same time, verify whether the selected questions cover all psychological dimensions. If the number of questions for a certain psychological dimension is insufficient, mark it as a potential missing dimension. S5-4. When the number of matched questions is insufficient or the questions for a certain psychological dimension are not fully covered, new questions are generated by calling the question template corresponding to the missing psychological dimension; during generation, the current psychological state label and the user's historical answer records are combined; the generated new questions are added to the dynamic question sequence. S5-5. Merge the matched questions and the newly generated questions, adjust the question order, and verify whether the length of the sequence and the coverage of the psychological dimension meet the assessment requirements, and finally generate a dynamic question sequence.

5. The online high-precision psychological assessment method based on big data according to claim 4, characterized in that, The process of adjusting the question order involves prioritizing questions that correspond to the current psychological state's psychological dimension, while inserting questions with low-relevance psychological dimensions between highly relevant questions. After each question is inserted, the difference in psychological dimension between it and the previous question is detected, and questions with psychological dimensions not yet touched in the user's answer record are inserted at the end of the sequence, resulting in a dynamic question sequence.

6. A high-precision online psychological assessment system based on big data, characterized in that: This system is used to implement the online high-precision psychological assessment method based on big data as described in any one of claims 1 to 5, comprising: The user terminal (1) is configured as follows: Obtain user information and transmit the user information to the server (2); Receive the initial question sequence generated by the server (2) and present it to the user; Collect user characteristic data during the question-answering process, and assess the user's psychological state based on the characteristic data; Transmit the user's psychological state to the server (2); The server-side (2) is configured as follows: Store historical assessment data and assessment questions; An initial question sequence is generated based on the user information, and the initial question sequence is transmitted to the user terminal (1). A dynamic question sequence is generated based on the user's psychological state, and the dynamic question sequence is transmitted to the user terminal (1).

7. The online high-precision psychological assessment system based on big data according to claim 6, characterized in that, The user terminal (1) includes: Display screen (11) is used to display assessment questions; The camera (12) is used to capture the user's facial features, acquire feature data, and connect to the first data processing unit (15) via an internal data bus. The information input module (13) is used to receive information input by the user and transmit it to the server (2) through the first network communication module (14). The first network communication module (14) is used to communicate with the server (2) via data. The first network communication module (14) and the server (2) are connected via a network. The first data processing unit (15) is configured to generate the user's psychological state based on the feature data during the user's answering process, and transmit the user's psychological state to the server (2) through the first network communication module (14). The server (2) includes: The database (21) is used to store historical assessment data and assessment questions, and is connected to the second data processing unit (23) via an internal data bus. The second network communication module (22) is used to communicate with the user terminal (1) via data. The second network communication module (22) is connected to the user terminal (1) via a network. The second data processing unit (23) includes: The initial question sequence generation module (151) is configured to filter assessment questions in the database (21) based on the user information and generate an initial question sequence; The dynamic question sequence generation module (152) is configured to generate a dynamic question sequence based on the user's psychological state.

8. The online high-precision psychological assessment system based on big data according to claim 7, characterized in that, The dynamic question sequence generation module (152) includes: The matching submodule is used to filter questions that match the psychological state tag from the database (21) based on the user's current psychological state, and add the selected questions to the dynamic question sequence; The coverage verification submodule is used to verify whether the selected questions cover all psychological dimensions. If it is found that the number of questions for a certain psychological dimension is insufficient, the psychological dimension is marked as a potential missing dimension. The new question generation submodule is used to generate new questions by calling the question template corresponding to the missing dimension when the number of matched questions is insufficient or the questions of a certain psychological dimension are not fully covered, and to supplement the dynamic question sequence with the new questions. The sorting submodule is used to sort the questions in the dynamic question sequence to ensure the progressiveness and logical coherence of the psychological dimension, and to verify whether the length of the sequence and the coverage of the psychological dimension meet the assessment requirements, and finally generate the dynamic question sequence.

Images