A mental health state prediction method and system based on correlation analysis

Abstract

The application belongs to the field of mental health assessment, and provides a mental health state prediction method and system based on association analysis, which comprises the following steps: performing privacy calculation based on federated learning to obtain mental assessment original data; preprocessing the mental assessment original data to obtain preprocessed mental assessment data; creating a two-dimensional storage matrix based on the preprocessed mental assessment data, and grouping the two-dimensional storage matrix to obtain a data set; performing association analysis based on the data set to construct a frequent tree, and obtaining a strong association rule table meeting a minimum support degree; selecting a characteristic dimension with strong correlation with other factors according to the strong association rule table to construct a mental characteristic, and using a trained mental health state prediction model to predict the mental health state. By changing the data set storage mode and scanning mode, the strong association rule can be obtained by scanning the data set only once, thereby saving the database storage space and improving the association rule algorithm mining efficiency.

Description

Technical Field

[0001] This invention belongs to the field of mental health assessment technology, specifically relating to a method and system for predicting mental health status based on association analysis. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Suicides due to mental health issues are increasingly common in today's society, making mental health a growing social concern. Due to a lack of awareness about mental illness and low willingness to seek treatment, timely detection and intervention for those with mental health problems is crucial. Currently used assessment methods suffer from limitations in data collection and statistical analysis, leading to discrepancies with reality. Furthermore, the results are limited to the psychological descriptions and interpretations of the test-takers, lacking predictive and intervention capabilities. In other words, existing methods only explain the results of assessments but cannot predict the future mental health of test-takers.

[0004] In terms of data acquisition, traditional methods rely on behavioral data from specific groups. The acquisition and use of this data can infringe on personal privacy to some extent, posing risks of privacy data exposure. Regarding data analysis, the FP-growth association rule algorithm is widely used. However, the traditional FP-growth algorithm compresses data records by constructing a tree structure. Mining frequent itemsets requires scanning the data records twice, and the constructed FP-tree is memory-intensive, consuming significant memory space and requiring improved efficiency. In data application, most online psychological assessment systems currently rely on scale assessments. While traditional scale assessments can determine whether an individual has certain psychological symptoms and their severity, they do not consider or can hardly predict an individual's mental health status. Summary of the Invention

[0005] To address the aforementioned issues, this invention proposes a method and system for predicting mental health status based on association analysis. While ensuring privacy, this invention guarantees the accuracy of psychological assessment data and provides it to users for big data analysis. By changing the data set storage and scanning methods, strong association rules can be obtained by scanning the data set only once, thus saving database storage space and improving the efficiency of association rule mining algorithms.

[0006] According to some embodiments, the first aspect of the present invention provides a method for predicting mental health status based on association analysis, employing the following technical solution:

[0007] A method for predicting mental health status based on association analysis, comprising:

[0008] Privacy-preserving computations are performed using federated learning to obtain raw data for psychological assessments.

[0009] The raw psychological assessment data is preprocessed to obtain preprocessed psychological assessment data.

[0010] A two-dimensional storage matrix is ​​created by scanning the preprocessed psychological assessment data, and the dataset is obtained by grouping the two-dimensional storage matrix.

[0011] Based on the dataset, a frequent tree is constructed for association analysis to obtain a table of strong association rules that meet the minimum support.

[0012] Based on the strong association rule table, psychological features are constructed by selecting feature dimensions that are strongly correlated with other factors, and the trained mental health status prediction model is used to predict mental health status.

[0013] According to some embodiments, a second aspect of the present invention provides a mental health status prediction system based on correlation analysis, employing the following technical solution:

[0014] A mental health status prediction system based on association analysis includes:

[0015] The data acquisition module is configured to perform privacy-preserving computations based on federated learning to obtain raw psychological assessment data.

[0016] The data preprocessing module is configured to preprocess the raw psychological test data to obtain preprocessed psychological test data.

[0017] The data grouping module is configured to scan and create a two-dimensional storage matrix based on preprocessed psychological assessment data, and then group the two-dimensional storage matrix to obtain the dataset.

[0018] The association analysis module is configured to construct a frequent tree based on the dataset to perform association analysis and obtain a table of strong association rules that meet the minimum support.

[0019] The psychological assessment module is configured to construct psychological features by selecting feature dimensions that are strongly correlated with other factors based on the strong association rule table, and to predict mental health status using a trained mental health status prediction model.

[0020] According to some embodiments, a third aspect of the present invention provides a computer-readable storage medium.

[0021] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a method for predicting mental health status based on association analysis as described in the first aspect above.

[0022] According to some embodiments, a fourth aspect of the present invention provides a computer device.

[0023] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the steps of a method for predicting mental health status based on association analysis as described in the first aspect above.

[0024] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0025] This invention analyzes and predicts the mental health status of test subjects by analyzing their raw psychological assessment data. It uses lateral federated learning technology to perform privacy-preserving computations on the psychological assessment data, achieving reliable computation and use of the data while protecting personal privacy. By utilizing the FP-growth improved algorithm and changing the data storage and scanning methods, strong association rules can be obtained by scanning the dataset only once, saving database storage space and improving the efficiency of association rule mining algorithms.

[0026] This invention provides a novel method for analyzing and predicting mental health status. By performing association rule analysis on psychological assessment data, a mental health status prediction model based on XG-Boost is constructed, outputting the probability that an individual's mental health status is abnormal, and accurately determining whether the individual has mental health problems.

[0027] This invention is based on a large amount of psychological assessment data generated by online assessments. It uses association rule mining technology to construct features and build a psychological health status prediction model, enabling the prediction of test takers' psychological state in advance. This helps psychologists to intervene in test takers with psychological abnormalities in a timely manner and provides new ideas for psychological researchers. Attached Figure Description

[0028] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0029] Figure 1 This is a flowchart of a method for predicting mental health status based on association analysis in an embodiment of the present invention;

[0030] Figure 2 This is a privacy computing logic architecture diagram of the psychological assessment system in this embodiment of the invention;

[0031] Figure 3 This is the improved FP-growth algorithm mining process in this embodiment of the invention. Detailed Implementation

[0032] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0033] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0034] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0035] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0036] Example 1

[0037] like Figure 1 As shown, this embodiment provides a method for predicting mental health status based on correlation analysis. This embodiment uses the application of this method to a server as an example for illustration. It is understood that this method can also be applied to terminals, and can also be applied to systems including terminals, servers, and other components, and can be implemented through interaction between the terminal and the server. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network servers, cloud communication, middleware services, domain name services, CDN security services, and big data and artificial intelligence platforms. The terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, etc., but is not limited to these. The terminal and server can be directly or indirectly connected via wired or wireless communication, which is not limited herein. In this embodiment, the method includes the following steps:

[0038] Privacy-preserving computations are performed using federated learning to obtain raw data for psychological assessments.

[0039] The raw psychological assessment data is preprocessed to obtain preprocessed psychological assessment data.

[0040] A two-dimensional storage matrix is ​​created by scanning the preprocessed psychological assessment data, and the dataset is obtained by grouping the two-dimensional storage matrix.

[0041] Based on the dataset, a frequent tree is constructed for association analysis to obtain a table of strong association rules that meet the minimum support.

[0042] Based on the strong association rule table, psychological features are constructed by selecting feature dimensions that are strongly correlated with other factors, and the trained mental health status prediction model is used to predict mental health status.

[0043] Specifically, the privacy-preserving computation based on federated learning to obtain the raw data for psychological assessment includes:

[0044] The system collects the initial mental health assessment results of multiple testers through the assessment system and uploads them to the server, providing a preliminary global model for the mental health assessment of the testers.

[0045] The evaluation system independently calculates model parameters locally using different sets of test personnel samples, encrypts the parameter information, and sends it to the server.

[0046] The server uses a weighted average algorithm based on homomorphic encryption to securely aggregate the encrypted parameter information, update the global model for the psychological health assessment of testers, and return the aggregated parameter information to the assessment system in an encrypted manner.

[0047] The evaluation system decrypts the encrypted and aggregated parameter information, uses the decrypted parameter information to update the local model, and enters the next round of training. This iterative cycle continues until the loss function converges.

[0048] The server aggregates the parameter information from the last round of local computation, outputs the model results, and generates the raw psychological assessment data after system privacy computation.

[0049] Specifically, the preprocessing of the raw psychological assessment data to obtain preprocessed psychological assessment data includes:

[0050] Remove non-critical information from the raw data of psychological assessments, retain fields that are strongly correlated with mental health status, and reduce the amount of data stored.

[0051] Missing values ​​are deleted or filled, and redundant data is removed;

[0052] The Chinese fields and factor scores were uniformly encoded to obtain the preprocessed psychological assessment data.

[0053] The process involves scanning and creating a two-dimensional storage matrix based on preprocessed psychological assessment data, and then grouping the two-dimensional storage matrix to obtain the dataset. Specifically:

[0054] The psychological assessment experimental dataset D and the minimum support count n are obtained based on the preprocessed psychological assessment data.

[0055] The first scan of the psychological test experimental dataset D begins. If the score of a factor is ≥2, then record 1 in the two-dimensional matrix of the psychological test data; otherwise, record 0.

[0056] Add a "sum" column to the last column of the two-dimensional matrix of psychological assessment data to count the number of records 1 in the current row;

[0057] Remove data from the two-dimensional matrix of psychological test data that do not meet the minimum support, and obtain a two-dimensional storage matrix that meets the minimum support n;

[0058] Based on the two-dimensional storage matrix that satisfies the minimum support n, the data set is scanned again in groups to obtain the dataset.

[0059] The dataset is obtained by further grouping and scanning based on the two-dimensional storage matrix that satisfies the minimum support n, including:

[0060] The scan is performed based on a two-dimensional storage matrix that satisfies the minimum support n. The n1 column is scanned. Since the value corresponding to m1n1 is 1, the scan continues downward until the value corresponding to the m row of that column is 0. The scan ends and a group (s1, s2, s3...) is established.

[0061] Continue scanning the next column. If there is a non-zero value in the column, continue scanning the non-zero positions until the column ends, and then continue scanning the next column. If the mean of the column is 0, automatically scan the next column until the second scan of the two-dimensional storage matrix that meets the minimum support n is completed.

[0062] After completing the scanning and grouping, the grouped dataset is obtained.

[0063] The process of constructing a frequent tree based on the dataset for association analysis yields a table of strong association rules that satisfy the minimum support, including:

[0064] Based on the dataset, a root node is created and inserted into the FP-Tree. If subsequent traversals are not empty, the relationship between the node and the node to be inserted is determined, thus completing the construction of the psychological assessment factor FP-Tree.

[0065] By calling the FP-growth function to mine frequent patterns in the FP-Tree, strong association rules between different dimensions that satisfy minimum confidence and minimum support are obtained in the psychological assessment experimental dataset D. The output is a table of strong association rules containing antecedent, consequent, support, and confidence.

[0066] The step of selecting feature dimensions with strong correlation to other factors to construct psychological features based on the strong correlation rule table, and using a trained mental health status prediction model to predict mental health status, includes:

[0067] Based on the obtained association rule table, data dimensionality is reduced by filtering low variance features, and features that are significantly different from mental health status are screened by the homogeneity of variance test. This yields feature dimensions that are strongly correlated with other factors, and constructs psychological features.

[0068] Based on psychological characteristics, a trained XG-Boost-based mental health prediction model is used to predict mental health status.

[0069] like Figure 1 As shown, the method described in this embodiment specifically includes:

[0070] First, using lateral federated learning technology, the raw data of the Symptom Checklist-90 (SCL90) assessment results were obtained after privacy computation. Second, data preprocessing operations such as data selection, data cleaning, data transformation, and data integration were performed on the raw data to construct the experimental dataset. Third, the FP-growth improved algorithm was used to mine frequent itemsets in the dataset to obtain strong association rules between the dimensions of the assessment scale. Finally, based on the association results, psychological dimension feature vectors were selected and constructed to build a mental health status prediction model based on XG-boost, and the predicted mental health status results were output.

[0071] The privacy computing logic architecture diagram of the psychological assessment system is as follows: Figure 2 As shown, data privacy computation based on federated learning:

[0072] To address mental health safety issues, this method employs federated learning technology for privacy protection. Taking students as an example, firstly, students A, B, through N complete the initial student mental health assessment through the assessment system and upload it to the server, providing a preliminary global model. Subsequently, the assessment system independently calculates model parameters (such as scores for a certain factor assessment result, average scores for a certain factor assessment, etc.) locally using different student sample sets {SA}, {SB}, ..., {Sn}, encrypts the parameter information, and then sends it to the server. At this point, the system server can use algorithms such as weighted averaging (gradient averaging, model averaging) based on homomorphic encryption to securely aggregate the sent parameter information, update the global model of student mental health assessment, and return the aggregated parameter information to the system in encrypted form. The system decrypts the received new parameter information and uses the decrypted parameter results to update the local model, entering the next round of training. This iterative cycle continues until the loss function converges. Finally, the server aggregates the parameter information calculated locally in the last round, outputs the model results, and generates the original psychological assessment data after the system's privacy calculations.

[0073] Data preprocessing:

[0074] The psychological assessment data collected using the lateral federated learning technique suffers from anomalies such as missing values, redundancy, and inconsistent formatting. This method addresses these issues through data selection, cleaning, and transformation. First, non-critical information such as gender and age is removed from the original data, while fields strongly correlated with mental health, such as somatization scores and obsessive-compulsive symptom scores, are retained to reduce data storage. Second, missing values ​​are deleted or filled, and redundant data is removed. Finally, to standardize the input format for subsequent model training data, Chinese language fields and factor scores are uniformly encoded.

[0075] Improved association rule mining:

[0076] To uncover strong associations between factors in mental health scales, traditional FP-growth association rule algorithms construct FP-Trees on each iteration, resulting in memory consumption, high memory usage, and low efficiency. This invention addresses the issue of array-based data of psychological assessment results by modifying the data storage and scanning methods during the data mining process. This effectively reduces the number of scans, thereby improving scanning efficiency and reducing storage space. The improved mining process is as follows: Figure 3 As shown.

[0077] (1) Constructing the dataset

[0078] First, input the psychological assessment experimental dataset D and the minimum support count n. Perform the first scan and store it as a two-dimensional matrix A. Refer to the scoring criteria of each item in the Symptom Checklist-90 (SCL-90) scale. If the score of a factor is ≥2, it is considered that there is an abnormal psychological state, and a 1 is recorded in the two-dimensional matrix A; otherwise, a 0 is recorded. Add a "sum" column to the last column of the psychological assessment data matrix A to count the number of problematic items in the row, i.e., the number of records of 1. Sort the matrix A in descending order according to the "sum" column. Now that the minimum support count is n, delete the data in the psychological assessment data matrix A that does not meet the minimum support, and obtain a two-dimensional storage matrix A1 that meets the minimum support n. Scan the newly generated two-dimensional storage matrix A1 and group it according to the experimental grouping requirements to obtain the grouped dataset S.

[0079] It should be noted that the psychological assessment experimental dataset D specifically refers to the preprocessed psychological assessment data; while the minimum support count is used to filter out data that does not meet the conditions, that is, data that meets the conditions is kept and data that does not meet the conditions is deleted, which is a step in the FP-growth algorithm process.

[0080] Grouping requirements: Scan the two-dimensional storage matrix A1 (m*n). For example, scan column n1. Since the value corresponding to m1n1 is 1, continue scanning downwards until the value corresponding to row m in that column is 0. The scan ends, and a group (s1, s2, s3...) is created. Continue scanning the next column. If a column contains non-zero values, continue scanning the non-zero positions until the column ends, then continue scanning the next column. If the mean of all values ​​in a column is 0, automatically scan the next column until the second scan of the two-dimensional storage matrix A1 is completed. After completing the scanning and grouping, the grouped dataset S is obtained.

[0081] (2) Constructing FP-Tree

[0082] Based on the grouped dataset S obtained from the above steps, a root node is created and inserted into the FP-Tree. If subsequent traversals are not empty, the relationship between this node and the node to be inserted is determined. It is checked whether the two nodes are parent-child; if so, the node is inserted directly, and the loop continues to the next node. If not, it is checked whether there is a shared ancestor; if so, and the two nodes have an inclusion relationship, the node is inserted; otherwise, the loop is exited. If none of the above conditions are met, a new branch is created, completing the construction of the psychological assessment factor FP-Tree.

[0083] (3) Mining of frequent items

[0084] The FP-Tree formed above is used to mine frequent patterns by calling the FP-growth function. First, mining begins from the leaf nodes upwards, identifying the item with the lowest support in the first frequent itemset and constructing its conditional pattern base. This then finds the set of prefix paths for the suffix patterns to be mined in the FP-Tree, forming a sub-database. Mining continues on this sub-database, constructing its conditional FP-Tree and recursively discovering frequent patterns within it. All subtrees are merged to obtain frequent itemsets between factors, which are stored in an association rule base. This yields strong association rules in the psychological assessment dataset D that satisfy minimum confidence and minimum support across different dimensions, outputting a table of strong association rules containing antecedent, consequent, support, and confidence.

[0085] Model building:

[0086] (1) Feature selection and construction

[0087] Based on the obtained association rule table, we selected feature dimensions with strong correlations to other factors to construct psychological features. Features with low variance and low or negative correlation with mental health status can affect the accuracy of model predictions. Therefore, we performed data dimensionality reduction by filtering low-variance features and used the homogeneity of variance test (F-test) to screen features that showed significant differences (P<0.05) with mental health status. Simultaneously, we used the diagnostic results of mental health status in the dataset as labels.

[0088] (2) Prediction model construction and training

[0089] An XG-Boost-based prediction model was constructed to predict mental health status. During model training, the constructed mental health features and labels were used as input. The dataset was partitioned as follows: 80% was randomly selected as the training set, and the remaining 20% ​​as the test set. A second 20% was randomly selected from the training set as the validation set, and 7-fold cross-validation was used to improve model robustness. For model evaluation, the optimal parameter combination was selected based on different evaluation metrics such as accuracy, recall, and F1 score.

[0090] (3) Model Output

[0091] The model outputs a floating-point number between [0, 1] as the probability of whether a mental health problem exists. When compared on the same dataset, it outperforms the XG-Boost prediction model based on the traditional FP-growth association rule algorithm.

[0092] Example 2

[0093] This embodiment provides a mental health status prediction system based on association analysis, including:

[0094] The data acquisition module is configured to perform privacy-preserving computations based on federated learning to obtain raw psychological assessment data.

[0095] The data preprocessing module is configured to preprocess the raw psychological test data to obtain preprocessed psychological test data.

[0096] The data grouping module is configured to scan and create a two-dimensional storage matrix based on preprocessed psychological assessment data, and then group the two-dimensional storage matrix to obtain the dataset.

[0097] The association analysis module is configured to construct a frequent tree based on the dataset to perform association analysis and obtain a table of strong association rules that meet the minimum support.

[0098] The psychological assessment module is configured to construct psychological features by selecting feature dimensions that are strongly correlated with other factors based on the strong association rule table, and to predict mental health status using a trained mental health status prediction model.

[0099] The examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the content disclosed in Embodiment 1 above. It should be noted that the above modules, as part of the system, can be executed in a computer system such as a set of computer-executable instructions.

[0100] The descriptions of each embodiment in the above embodiments have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0101] The proposed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and the division of modules described above is only a logical functional division. In actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed.

[0102] Example 3

[0103] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the association analysis-based mental health status prediction method described in Embodiment 1 above.

[0104] Example 4

[0105] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the association analysis-based mental health status prediction method described in Embodiment 1 above.

[0106] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0107] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0108] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0109] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0110] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0111] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for predicting mental health status based on association analysis, characterized in that, include: Privacy-preserving computation based on federated learning yields raw psychological assessment data. This process includes: acquiring initial psychological health assessment results from multiple test takers via an assessment system and uploading them to a server, providing a preliminary global model for the psychological health assessment; the assessment system independently calculates model parameters locally using different test taker sample sets, encrypts the parameter information, and sends it to the server; the server uses a weighted average algorithm based on homomorphic encryption to securely aggregate the encrypted parameter information, updates the global psychological health assessment model, and returns the aggregated parameter information to the assessment system in encrypted form; the assessment system decrypts the encrypted aggregated parameter information and uses the decrypted parameter information to update its local model, entering the next round of training, iterating in this cycle until the loss function converges; the server aggregates the parameter information calculated locally in the final round, outputs the model results, and generates the raw psychological assessment data after system privacy computation. The raw psychological assessment data is preprocessed to obtain preprocessed psychological assessment data. A two-dimensional storage matrix is ​​created by scanning the preprocessed psychological assessment data, and the two-dimensional storage matrix is ​​then grouped to obtain the dataset. This process includes: obtaining a psychological assessment experimental dataset D and a minimum support count of n based on the preprocessed psychological assessment data; starting the first scan of the psychological assessment experimental dataset D, recording 1 in the two-dimensional matrix if the factor score is ≥2, otherwise recording 0; adding a "sum" column to the last column of the two-dimensional matrix to count the number of records of 1 in the current row; deleting data from the two-dimensional matrix that does not meet the minimum support, resulting in a two-dimensional storage matrix that meets the minimum support n; and performing another grouping scan based on the two-dimensional storage matrix that meets the minimum support n to obtain the dataset. Based on the dataset, a frequent tree is constructed for association analysis to obtain a table of strong association rules that meet the minimum support. Based on the strong association rule table, psychological features are constructed by selecting feature dimensions that are strongly correlated with other factors, and the trained mental health status prediction model is used to predict mental health status.

2. The method for predicting mental health status based on association analysis as described in claim 1, characterized in that, The preprocessing of the raw psychological assessment data to obtain preprocessed psychological assessment data includes: Remove non-critical information from the raw data of psychological assessments, retain fields that are strongly correlated with mental health status, and reduce the amount of data stored. Missing values ​​are deleted or filled, and redundant data is removed; The Chinese fields and factor scores were uniformly encoded to obtain the preprocessed psychological assessment data.

3. The method for predicting mental health status based on association analysis as described in claim 1, characterized in that, The dataset is obtained by further grouping and scanning based on the two-dimensional storage matrix that satisfies the minimum support n, including: The scan is performed based on a two-dimensional storage matrix that satisfies the minimum support n. The n1 column is scanned. Since the value corresponding to m1n1 is 1, the scan continues downward until the value corresponding to the m row of that column is 0. The scan ends and a group (s1, s2, s3...) is established. Continue scanning the next column. If there is a non-zero value in the column, continue scanning the non-zero positions until the column ends, and then continue scanning the next column. If the mean of the column is 0, automatically scan the next column until the second scan of the two-dimensional storage matrix that meets the minimum support n is completed. After completing the scanning and grouping, the grouped dataset is obtained.

4. The method for predicting mental health status based on association analysis as described in claim 1, characterized in that, The process of constructing a frequent tree based on the dataset for association analysis yields a table of strong association rules that satisfy the minimum support, including: Based on the dataset, a root node is created and inserted into the FP-Tree. If subsequent traversals are not empty, the relationship between the current node and the node to be inserted is determined, thus completing the construction of the psychological assessment factor FP-Tree. By calling the FP-growth function to mine frequent patterns in the FP-Tree, strong association rules between different dimensions that satisfy minimum confidence and minimum support are obtained in the psychological assessment experimental dataset D. The output is a table of strong association rules containing antecedent, consequent, support, and confidence.

5. The method for predicting mental health status based on association analysis as described in claim 1, characterized in that, The step of selecting feature dimensions with strong correlation to other factors to construct psychological features based on the strong correlation rule table, and using a trained mental health status prediction model to predict mental health status, includes: Based on the obtained association rule table, data dimensionality is reduced by filtering low variance features, and features that are significantly different from mental health status are screened by the homogeneity of variance test. Feature dimensions that are strongly correlated with other factors are obtained, and psychological features are constructed. Based on psychological characteristics, a trained XG-Boost-based mental health prediction model is used to predict mental health status.

6. A mental health status prediction system based on association analysis, characterized in that, include: The data acquisition module is configured to perform privacy-preserving computation based on federated learning to obtain raw psychological assessment data. This includes: acquiring initial psychological health assessment results from multiple test takers completed through the assessment system and uploading them to the server, providing a preliminary global model for the psychological health assessment; the assessment system independently calculates model parameters locally using different test taker sample sets, encrypts the parameter information, and sends it to the server; the server uses a weighted average algorithm based on homomorphic encryption to securely aggregate the encrypted parameter information, updates the global model for the psychological health assessment, and returns the aggregated parameter information to the assessment system in encrypted form; the assessment system decrypts the encrypted and aggregated parameter information, uses the decrypted parameter information to update the local model, and enters the next round of training, iterating in this loop until the loss function converges; the server aggregates the parameter information calculated locally in the last round, outputs the model results, and generates the raw psychological assessment data after system privacy computation. The data preprocessing module is configured to preprocess the raw psychological test data to obtain preprocessed psychological test data. The data grouping module is configured to scan and create a two-dimensional storage matrix based on preprocessed psychological assessment data, and then group the two-dimensional storage matrix to obtain datasets. This includes: obtaining a psychological assessment experimental dataset D and a minimum support count of n based on the preprocessed psychological assessment data; performing the first scan on the psychological assessment experimental dataset D, recording 1 in the two-dimensional matrix if the factor score is ≥2, otherwise recording 0; adding a "sum" column to the last column of the two-dimensional matrix to count the number of records of 1 in the current row; deleting data from the two-dimensional matrix that does not meet the minimum support, resulting in a two-dimensional storage matrix that meets the minimum support n; and performing another grouping scan based on the two-dimensional storage matrix that meets the minimum support n to obtain the datasets. The association analysis module is configured to construct a frequent tree based on the dataset to perform association analysis and obtain a table of strong association rules that meet the minimum support. The psychological assessment module is configured to construct psychological features by selecting feature dimensions that are strongly correlated with other factors based on the strong association rule table, and to predict mental health status using a trained mental health status prediction model.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the mental health status prediction method based on association analysis as described in any one of claims 1-5.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the mental health status prediction method based on association analysis as described in any one of claims 1-5.

Images