A student intelligent psychological state evaluation system

Abstract

The application discloses a student intelligent psychological state evaluation system, and relates to the technical field of data communication. First, the quality of student published data is checked and automatically corrected to ensure the reliability of subsequent processing input. Then, data sequence clusters are divided through clustering analysis, and resource scheduling strategies are dynamically adjusted based on the similarity confidence of each cluster to realize the optimization of analysis frequency and sampling rate of different clusters. Meanwhile, the resource scheduling strategy in the cluster is dynamically adjusted, and the optimal task allocation plan is intelligently calculated and issued according to the scheduling exposure index and resource proportion index of each sequence cluster to realize the fine matching of computing power and manpower. Finally, the real-time monitoring of key performance indicators improves the real-time performance and accuracy of student psychological state evaluation, and significantly optimizes the resource utilization efficiency and stability of the system in a high-concurrency scenario.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data communication, and particularly relates to a student intelligent psychological state evaluation system. BACKGROUND

[0002] The student intelligent psychological state evaluation system realizes real-time monitoring and evaluation of the psychological state of students by integrating multi-modal data acquisition, intelligent analysis and dynamic scheduling function modules.

[0003] For example, the disclosure number CN119626463A discloses a student mental health early warning method and early warning system based on AI empowerment. By constructing an event listener and an event trigger, the metadata of student mental health activities is collected in real time. The psychological activity state of the student when performing structured mental health activities is converted into a difference equation. The psychological activity state difference equation is solved with metadata to obtain the fluctuation state matrix of the student's psychological activity at each metadata generation. The fluctuation state matrix is clustered using the associated crisis rating heat map of the student's mental health to train the rating model and generate the mental health crisis rating model. The standardized mental health rating model and the mental health crisis rating model of the student's psychological state in the healthy standard state are mapped and the distribution distance offset is analyzed for monitoring and early warning.

[0004] For example, the announcement number CN118380091B discloses an intelligent data management system based on a psychological evaluation instrument. The system includes a psychological test information collection module that collects physiological and behavioral data of students through a psychological evaluation instrument. A student physiological index analysis module analyzes the data to obtain a physiological state coincidence index. A student reaction index analysis module evaluates a reaction state coincidence index based on behavioral data. A psychological state comprehensive analysis module combines the two to calculate a student's mental health index and stores it in an evaluation information database. In addition, a guidance effect evaluation feedback module extracts historical data from the database to evaluate the psychological guidance effect and feeds back to school administrators. The evaluation information database also stores physiological index reference values, facial reaction evaluation weights, facial image sets, and language reaction evaluation weights.

[0005] The above-mentioned technology at least has the following technical problems:

[0006] In the evaluation of students' intelligent psychological state, the quality of the data published by students is uneven, there is a lack of systematic quality evaluation and correction for multi-modal input, and in the task allocation process of data processing, there may be uneven resource allocation, response time delay, static or polling scheduling in task allocation, etc., resulting in waste of computing power and manpower, response delay and priority misplacement, and the ability to couple the propagation or resource index into abnormal detection is also lacking, so that early warning and automatic load reduction cannot be achieved when the load surges or heat sources concentrate, which may further lead to delayed identification of student psychological abnormalities and inaccurate intervention measures, thereby affecting the effectiveness of student mental health management. SUMMARY

[0007] In view of the deficiencies of the prior art, the student intelligent psychological state evaluation system solves the problems in the above background art.

[0008] To achieve the above purpose, the following technical solutions are adopted: a student intelligent psychological state evaluation system, comprising a data analysis and correction module, configured to obtain student published data in an AI tree hole for preliminary analysis, obtain a student published data quality compliance representation value, and correct the student published data according to the student published data quality compliance representation value.

[0009] A clustering analysis module is configured to extract the corrected student published data and perform clustering analysis to obtain a clustering similarity confidence of the student published data, filter the student published data sequence cluster according to the clustering similarity confidence of the student published data, and dynamically adjust the student published data processing resources.

[0010] A task allocation and scheduling module is configured to analyze the scheduling data of the student published data sequence cluster to obtain a scheduling exposure index of the student published data sequence cluster, and perform task allocation and scheduling of the student published data according to the scheduling exposure index of the student published data sequence cluster.

[0011] A task allocation and warning module is configured to obtain key index data in the task allocation process for analysis, obtain a task allocation process abnormality index in combination with the scheduling exposure index of the student published data sequence cluster, screen according to the task allocation process abnormality index, and perform early warning on the task allocation process.

[0012] Further, a student published data quality compliance representation value is obtained, and the specific process is as follows: student published data in the AI tree hole is extracted, including picture peak signal-to-noise ratio, contrast, sentence structure integrity rate, and redundant word quantity; the picture peak signal-to-noise ratio is compared with the defined picture peak signal-to-noise ratio, the contrast is compared with the defined contrast, the sentence structure integrity rate is compared with the defined sentence structure integrity rate, and the redundant word quantity is compared with the defined redundant word quantity; after introducing a weight coefficient, coupling is performed to obtain the student published data quality compliance representation value, which is used to evaluate the information reliability of the student published data.

[0013] Further, the student published data is corrected according to the student published data quality compliance representation value, and the specific process is as follows: a monitoring time period is preset, the student published data quality compliance representation value is extracted in the monitoring time period, and is compared with a set student published data quality compliance representation threshold value; if the student published data quality compliance representation value is higher than or equal to the student published data quality compliance representation threshold value, first correction is performed; if the student published data quality compliance representation value is lower than the student published data quality compliance representation threshold value, second correction is performed.

[0014] Further, the student published data is filtered according to the clustering similarity confidence of the student published data, and a student published data sequence cluster is obtained, and the specific process is as follows: the corrected student published data is extracted and clustering analysis is performed to obtain the clustering similarity confidence of the student published data; according to the clustering similarity confidence of the student published data, the student published data sequence cluster is obtained, which includes a student published first data sequence cluster and a student published second data sequence cluster.

[0015] Further, the student published data processing resource is dynamically adjusted, and the specific process is as follows: the student published first data sequence cluster and the student published second data sequence cluster are extracted; the real-time analysis frequency of the student published data processing resource in the student published first data sequence cluster is dynamically adjusted; and the data sampling rate of the student published data processing resource in the student published second data sequence cluster is dynamically adjusted.

[0016] Further, a student published data sequence cluster scheduling exposure index is obtained, and the specific process is as follows: in the monitoring time period, the server cumulative throughput proportion, the memory cumulative residence time, and the cumulative propagation times corresponding to the student published data are obtained; the server cumulative throughput proportion corresponding to the student published data is compared with the defined server resource cumulative throughput proportion, the memory cumulative residence time is compared with the defined memory cumulative residence time, and the cumulative propagation times are compared with the defined propagation times; after introducing a weight coefficient, coupling is performed to obtain the student published data sequence cluster scheduling exposure index, which is used to evaluate the occupation degree of the data sequence cluster to the system resource.

[0017] Further, the published data of the student is scheduled and distributed according to the scheduling exposure index of the published data sequence cluster of the student, and the specific process is that the scheduling exposure index of the published data sequence cluster of the student is extracted, and is compared with the scheduling exposure index threshold of the published data sequence cluster of the student, if the scheduling exposure index of the published data sequence cluster of the student is higher than or equal to the scheduling exposure index threshold of the published data sequence cluster of the student, the content corresponding to the scheduling exposure index of the published data sequence cluster of the student is preferentially distributed to the real-time processing queue.

[0018] Further, the task distribution process abnormality index is obtained, and the specific process is that the key indicator data in the task distribution process is extracted in the monitoring time period, including the task distribution success rate, the task distribution response time and the task completion rate, the task distribution success rate is compared with the defined task distribution success rate, the task distribution response time is compared with the defined task distribution response time, and the task completion rate is compared with the defined task completion rate, and the scheduling exposure index of the published data sequence cluster of the student and the weight coefficient are coupled to obtain the task distribution process abnormality index, the task distribution process abnormality index is used to evaluate the stability of the task distribution mechanism.

[0019] Further, screening is performed according to the task distribution process abnormality index, and the specific process is that a monitoring period is preset, and the task distribution process abnormality index in each monitoring time period is obtained in the monitoring period, and the number of monitoring time periods with the task distribution process abnormality index higher than or equal to the task distribution process abnormality index threshold is counted.

[0020] Further, the task distribution process is warned, and the specific process is that the number of monitoring time periods with the task distribution process abnormality index higher than or equal to the task distribution process abnormality index threshold is extracted in each monitoring time period, and is recorded as the buffer number, if the buffer number is higher than the number of preset monitoring time periods, the task distribution process is warned.

[0021] The one or more technical solutions provided in the embodiments of the application have at least the following technical effects or advantages:

[0022] (1) The student intelligent psychological state evaluation system provided by the application first performs quality verification and automatic correction on the published data of the student, ensures the reliability of subsequent feature extraction, then divides data sequence clusters through clustering analysis, dynamically adjusts the cluster resource scheduling strategy, and then intelligently calculates and issues the optimal task distribution plan according to the scheduling exposure index and resource proportion index of each sequence cluster, realizes fine matching of computing power and manpower, and finally, the key performance indicators are monitored in real time, which improves the real-time and accuracy of the student psychological state evaluation, and significantly optimizes the resource utilization efficiency and stability of the system in a high-concurrency scenario.

[0023] (2) The application obtains a student published data quality meets a representation value, and triggers two correction strategies in a preset monitoring window according to the representation value, screens out a first data sequence cluster and a second data sequence cluster, dynamically optimizes a resource scheduling strategy at a cluster level granularity to ensure basic reliability of input data, effectively solves problems such as uneven multi-modal data quality, inaccurate clustering and grouping, static and inefficient resource scheduling, and lagging early warning mechanism, and significantly improves data processing reliability, task allocation accuracy and system stability of the student intelligent psychological state evaluation system in a high concurrency environment.

[0024] (3) The application obtains a scheduling exposure index, helps to preferentially allocate corresponding content of the student published data to a real-time processing queue, ensures that content with high resource occupation and large influence is preferentially processed, and thus improves response efficiency and resource utilization of the system.

[0025] (4) The application obtains a task allocation process abnormality index by continuously monitoring key indicators of the task allocation process, effectively avoids task delay or loss caused by allocation imbalance, helps the system to discover abnormal conditions in time, gives early warning, enhances sensitivity and response ability of the system to abnormal conditions, and ensures stability and efficiency of the system. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 is a schematic diagram of a system module of the application;

[0027] Figure 2 is a schematic diagram of a logic flow of the student intelligent psychological state evaluation system of the application;

[0028] Figure 3 is a schematic diagram of a logic flow of the student published data processing resource dynamic adjustment of the application. DETAILED DESCRIPTION

[0029] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0030] In the description of the application, it should be understood that the terms "opening", "upper", "lower", "thickness", "top", "middle", "length", "inner", "periphery" and the like indicate the orientation or positional relationship, which are only for the convenience of describing the application and simplifying the description, and do not indicate or imply that the indicated components or elements must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation of the application.

[0031] Referring to Figure 1 The embodiment of the present application provides a technical scheme: a student intelligent psychological state evaluation system, comprising a data analysis correction module, which is used for obtaining student published data in an AI tree hole for preliminary analysis to obtain a student published data quality compliance representation value, and correcting the student published data according to the student published data quality compliance representation value.

[0032] A clustering analysis module is used for extracting the corrected student published data and performing clustering analysis to obtain a clustering similarity confidence of the student published data, screening according to the clustering similarity confidence of the student published data, obtaining a student published data sequence cluster, and dynamically adjusting a student published data processing resource.

[0033] A task allocation and scheduling module is used for analyzing scheduling data of the student published data sequence cluster to obtain a scheduling exposure index of the student published data sequence cluster, and performing task allocation and scheduling on the student published data according to the scheduling exposure index of the student published data sequence cluster.

[0034] A task allocation and early warning module is used for obtaining key indicator data in a task allocation process for analysis, obtaining a task allocation process abnormality index in combination with the scheduling exposure index of the student published data sequence cluster, screening according to the task allocation process abnormality index, and performing early warning on the task allocation process.

[0035] Specifically, the student published data quality compliance representation value is obtained, and the specific process is as follows: student published data in an AI tree hole is extracted, including a picture peak signal-to-noise ratio, a contrast, a sentence structure integrity rate, and a redundant word quantity, the picture peak signal-to-noise ratio and a picture defined peak signal-to-noise ratio, the contrast and a defined contrast, the sentence structure integrity rate and a defined sentence structure integrity rate, and the redundant word quantity and a defined redundant word quantity are respectively analyzed in terms of proportion, and after introducing a weight coefficient, coupling is performed to obtain the student published data quality compliance representation value, which is used to evaluate the information reliability of the student published data.

[0036] It should be noted that the student published data quality compliance representation value is specifically analyzed under the following conditions: ;

[0037] In the formula, represents the student published data quality compliance representation value represents the picture peak signal-to-noise ratio, represents the set picture defined peak signal-to-noise ratio, represents the contrast, represents the defined contrast, represents the sentence structure integrity rate, represents the defined sentence structure integrity rate, represents the number of redundant repetitions, represents the number of redundant repetitions, represents the weight coefficient corresponding to the set picture peak signal-to-noise ratio, represents the weight coefficient corresponding to the set contrast, represents the weight coefficient corresponding to the set sentence structure integrity rate, represents the weight coefficient corresponding to the set number of redundant repetitions.

[0038] It should be noted that in the data content quality evaluation, the picture peak signal-to-noise ratio, the contrast, the sentence structure integrity rate and the number of redundant repetitions show close interaction: when the picture peak signal-to-noise ratio is low, it means that the distinction between background noise and signal decreases, directly compressing the effective contrast of the picture, making it difficult to distinguish the details of the edge; once the contrast is limited, the text extraction of high-frequency information in words and images will be missed, thus reducing the sentence structure integrity rate; at the same time, if the sentence structure integrity rate decreases, the system often compensates for the content loss by generating excessive prompts or placeholders, resulting in a sharp increase in the number of redundant repetitions, and a large number of repeated words will further interfere with the context association of the natural language understanding model, making the sentence parsing efficiency and accuracy double decline; the increase of redundant repetitions will trigger the image and text quality evaluation module to re-audit the original material, when detecting text-level problems, the algorithm will reduce the joint confidence of the image and text of the message, and thus affect the final automatic intervention decision; therefore, in practical application, the peak signal-to-noise ratio and contrast must be optimized simultaneously, and the number of redundant repetitions must be strictly controlled in the back-end natural language processing stage, in order to maintain a high level of sentence structure integrity rate, so as to realize the coordinated improvement of visual and text analysis and the reliable stability of the overall performance of the system.

[0039] It should be noted that the weighting coefficients for image peak signal-to-noise ratio (PSNR), contrast ratio, sentence structure integrity rate, and redundant word count are typically set between 0 and 1. For example, by constructing a mapping table between PSNR and weighting coefficients, the real-time detected PSNR is input into the database to quickly obtain the corresponding weighting coefficient. Similarly, for contrast ratio, a mapping table is constructed between contrast ratio and weighting coefficients, and the real-time detected contrast ratio is input into the database to quickly obtain the corresponding weighting coefficient. Likewise, for sentence structure integrity rate, a pre-established mapping table is used, and the real-time measured sentence structure integrity rate is input into the database to quickly obtain the corresponding weighting coefficient. Likewise, for redundant word count, a pre-established mapping table is used, and the real-time measured redundant word count is input into the database to quickly obtain the corresponding weighting coefficient.

[0040] Specifically, the student publication data is corrected based on the student publication data quality compliance characterization value. The specific process is as follows: a monitoring period is preset, the student publication data quality compliance characterization value is extracted during the monitoring period, and compared with the set student publication data quality compliance characterization threshold. If the student publication data quality compliance characterization value is higher than or equal to the student publication data quality compliance characterization threshold, the first correction is performed. If the student publication data quality compliance characterization value is lower than the student publication data quality compliance characterization threshold, the second correction is performed.

[0041] It should be noted that the specific process of the first correction is as follows: When performing the first correction on text data, spelling errors and punctuation problems are corrected using an automatic grammar checker; when performing the second correction, the sentence logic is reconstructed and missing data is supplemented by combining a semantic analysis model; when performing the first correction on image data, redundant areas are cropped, and when performing the second correction on image data, super-resolution reconstruction technology is used to improve the texture details of low-resolution images.

[0042] Specifically, the student publication data sequence clusters are obtained by filtering based on the cluster similarity confidence of the student publication data. The specific process is as follows: extract the corrected student publication data and perform cluster analysis to obtain the cluster similarity confidence of the student publication data. Based on the cluster similarity confidence of the student publication data, the student publication data sequence clusters are obtained. The student publication data sequence clusters include the first student publication data sequence cluster and the second student publication data sequence cluster.

[0043] It should be noted that the student posting data sequence clusters refer to the automatic aggregation of all posts into several groups with related topics and consistent data quality, based on the similarity between the corrected student posts and the merged quality content.

[0044] It should be noted that dividing the student publication data sequence clusters into a first data sequence cluster and a second data sequence cluster allows the system to dynamically adjust resource allocation and processing priorities based on the data. For the first data sequence cluster, which contains data with highly consistent quality characteristics and high system resource consumption, the sampling rate and analysis frequency can be increased, facilitating timely detection and intervention of potential psychological problems in subsequent systems. For the second data sequence cluster, the system can reduce the sampling rate, saving resources. This not only improves the system's efficiency and response speed but also provides a data foundation for future mental health assessments.

[0045] It should be noted that the specific process for obtaining the cluster similarity confidence of student publication data is as follows: For each corrected student publication data, its similarity with all other data samples in the quality dimension is calculated. For example, a quality feature vector is generated using four values: peak signal-to-noise ratio, contrast, text structure integrity rate, and redundancy rate. Then, the Euclidean distance between the quality feature vectors of any two publications is calculated to measure the similarity between the two publications in terms of quality and content. Next, the quality similarity vector and content similarity vector of each data point are input into a deep multimodal fusion network to obtain a joint feature vector containing information on both quality and content. Finally, based on these fusion vectors, the system uses an improved weighted K-means clustering algorithm—automatically dividing the data into several clusters under the metric of weighted Euclidean distance—and uses the normalized distance value from each data point to the center of its cluster as its cluster similarity confidence, so that each published data sequence cluster generated in the end maintains a high degree of consistency in both data quality and content theme.

[0046] Specifically, the student's publication data processing resources are dynamically adjusted. The specific process is as follows: extract the student's first publication data sequence cluster and the student's second publication data sequence cluster; dynamically adjust the real-time analysis frequency of the student's publication data processing resources in the student's first publication data sequence cluster; and dynamically adjust the data sampling rate of the student's publication data processing resources in the student's second publication data sequence cluster.

[0047] It should be noted that for the first data sequence cluster, the real-time analysis frequency is dynamically adjusted based on the student publication data quality compliance characterization value. For example, when the student publication data quality compliance characterization value exceeds the preset threshold, the analysis frequency is increased by 50% from the baseline value. For the second data sequence cluster, the data sampling rate is adjusted, and the sampling rate is compressed from 100% to 70% to reduce redundant calculations.

[0048] Specifically, the scheduling exposure index of the student's published data sequence cluster is obtained through the following process: during the monitoring period, the cumulative server throughput ratio, cumulative memory dwell time, and cumulative propagation times corresponding to the student's published data are obtained. The cumulative server throughput ratio corresponding to the student's published data is then compared with the defined cumulative throughput ratio of server resources, the cumulative memory dwell time is compared with the defined cumulative memory dwell time, and the cumulative propagation times are compared with the defined propagation times. After introducing weight coefficients, these are coupled to obtain the scheduling exposure index of the student's published data sequence cluster. The scheduling exposure index of the student's published data sequence cluster is used to evaluate the degree of system resource occupation by the data sequence cluster.

[0049] It should be noted that the scheduling exposure index of the student's published data sequence cluster is analyzed under the following conditions:

[0050] ;

[0051] ;

[0052] ;

[0053] in, This represents the scheduling exposure index of the student's published data sequence cluster. This represents the scheduling exposure index of the student's first published data sequence cluster. This represents the scheduling exposure index of the student's published second data sequence cluster. This represents the weighting coefficient corresponding to the scheduling exposure index of the first data sequence cluster. This represents the weighting coefficient corresponding to the scheduling exposure index of the set second data sequence cluster. This represents the cumulative throughput percentage of server resources for students publishing high-exposure content in the first data sequence cluster. This represents the cumulative throughput percentage of server resources used to represent the high-exposure content of the second data sequence cluster published by students. This indicates the cumulative throughput percentage defined by the set server resources. This represents the cumulative memory residency time of high-exposure content from the first data sequence cluster published by a student. This represents the cumulative memory residency time of high-exposure content from the second data sequence cluster published by students. This indicates the defined cumulative memory residency time. The cumulative number of times high-exposure content from students' first data sequence clusters has been disseminated. The cumulative number of times students publish highly exposed content from the second data sequence cluster has been disseminated. This indicates the set number of propagation attempts. This indicates the weighting coefficient corresponding to the cumulative throughput percentage of the set server resources. This represents the weighting coefficient corresponding to the set cumulative memory residency time. This represents the weighting coefficient corresponding to the set cumulative propagation count.

[0054] It should be noted that the cumulative throughput of server resources for high-exposure content, the cumulative memory dwell time, and the cumulative number of disseminations are closely intertwined. Once the cumulative throughput of high-exposure content continues to rise, it will increase the request concurrency of backend services, thereby increasing the cumulative memory dwell time. As the dwell time lengthens, memory pressure increases, which in turn leads to a short-term decrease in throughput and a longer request response time. This delay may, in turn, cause a sudden increase in the cumulative number of disseminations, because multiple repeated requests caused by client retries or queuing further increase the overall throughput demand of the server.

[0055] It should be noted that the weight coefficients corresponding to the scheduling exposure index of the first data sequence cluster, the scheduling exposure index of the second data sequence cluster, the cumulative throughput ratio of server resources, the cumulative memory dwell time, and the cumulative propagation times are usually set between 0 and 1. For example, by constructing mapping tables between the scheduling exposure index of the first data sequence cluster and the scheduling exposure index of the second data sequence cluster and their corresponding weight coefficients, and inputting the real-time obtained scheduling exposure indexes of the first and second data sequence clusters into the corresponding mapping tables in the database, the weight coefficients corresponding to the scheduling exposure indexes of the first and second data sequence clusters can be quickly obtained. Similarly, for the cumulative server resource throughput ratio, cumulative memory dwell time, and cumulative propagation times of high-exposure content, by constructing mapping tables between the cumulative server resource throughput ratio, cumulative memory dwell time, and cumulative propagation times of high-exposure content and their corresponding weight coefficients, and inputting the real-time detected cumulative server resource throughput ratio, cumulative memory dwell time, and cumulative propagation times of high-exposure content into the mapping tables in the database, the weight coefficients corresponding to the cumulative server resource throughput ratio, cumulative memory dwell time, and cumulative propagation times of high-exposure content can be quickly obtained.

[0056] It should be noted that high-exposure content refers to content that meets user interaction metrics such as a click-through rate exceeding a set threshold or a dwell time exceeding a set threshold.

[0057] It should be noted that the high-exposure content of the first data sequence cluster will be prioritized for manual review through full-volume in-depth analysis. The second data sequence cluster, on the other hand, will be reviewed using local hotspot sampling, which only samples and reviews related high-exposure content, helping to avoid data interference from discrete data.

[0058] Specifically, the task allocation and scheduling of students' published data is performed based on the scheduling exposure index of the students' published data sequence clusters. The specific process is as follows: extract the scheduling exposure index of the students' published data sequence clusters and compare it with the set scheduling exposure index threshold of the students' published data sequence clusters. If the scheduling exposure index of the students' published data sequence clusters is higher than or equal to the scheduling exposure index threshold of the students' published data sequence clusters, then the content corresponding to the scheduling exposure index of the students' published data sequence clusters is preferentially allocated to the real-time processing queue.

[0059] It should be noted that prioritizing the allocation of content corresponding to the exposure index of students' published data sequence clusters to the real-time processing queue helps ensure the timeliness of high-priority tasks and optimize system resource allocation.

[0060] Specifically, the task allocation process anomaly index is obtained through the following process: key indicator data in the task allocation process are extracted during the monitoring period, including task allocation success rate, task allocation response time, and task completion rate. The task allocation success rate is compared with the defined task allocation success rate, the task allocation response time is compared with the defined task allocation response time, and the task completion rate is compared with the defined task completion rate. The scheduling exposure index and weighting coefficient of the student's published data sequence cluster are then introduced and coupled to obtain the task allocation process anomaly index. The task allocation process anomaly index is used to evaluate the stability of the task allocation mechanism.

[0061] It should be noted that the specific analysis conditions for the task allocation process anomaly index are as follows:

[0062] ;

[0063] In the formula, Indicates the abnormality index of the task allocation process. This indicates the success rate of task assignment. This indicates the defined success rate of task assignment. Indicates the task allocation response time. This indicates the defined response time for the assigned task. Indicates the task completion rate. This indicates the task completion rate. This represents the scheduling exposure index of the student's published data sequence cluster. This represents the weighting coefficient corresponding to the set task allocation success rate. This represents the weighting coefficient corresponding to the set task allocation response time. This represents the weighting coefficient corresponding to the set task completion rate. This represents the weighting coefficient corresponding to the scheduling exposure index of the student's published data sequence cluster.

[0064] It should be noted that there is a close coupling between the task allocation success rate, task allocation response time, task completion rate, and the scheduling exposure index of the student published data sequence cluster: when the task allocation success rate decreases, it usually means that the backend service is failing to process under high concurrency or is being rate-limited. At this time, the response time will often be extended, which will lead to a surge in the cumulative number of system retries, which may push up the scheduling exposure index of the corresponding cluster. Once the exposure index rises, the backend must retain more session and state data, which will lead to the shortage of memory and computing resources, and the task completion rate may drop.

[0065] It should be noted that the weighting coefficients corresponding to task allocation success rate, task allocation response time, task completion rate, and the scheduling exposure index of student published data sequence clusters are typically set between 0 and 1. For example, by constructing mapping tables between task allocation success rate, task allocation response time, and task completion rate and their corresponding weighting coefficients, and inputting the real-time obtained task allocation success rate, task allocation response time, and task completion rate into the corresponding mapping tables in the database, the weighting coefficients corresponding to task allocation success rate, task allocation response time, and task completion rate can be quickly obtained. Similarly, for the scheduling exposure index of student published data sequence clusters, by constructing a mapping table between the scheduling exposure index of student published data sequence clusters and its corresponding weighting coefficients, and inputting the real-time obtained scheduling exposure index of student published data sequence clusters into the corresponding mapping table in the database, the weighting coefficient corresponding to the scheduling exposure index of student published data sequence clusters can be quickly obtained.

[0066] Specifically, the screening is based on the task allocation process abnormality index. The specific process is as follows: a monitoring period is preset, the task allocation process abnormality index is obtained in each monitoring time period within the monitoring period, and the number of monitoring time periods in which the task allocation process abnormality index is higher than or equal to the task allocation process abnormality index threshold is counted.

[0067] Specifically, the task allocation process is given an early warning. The process is as follows: extract the number of monitoring time periods in each monitoring time period where the abnormal index of the task allocation process is higher than or equal to the threshold of the abnormal index of the task allocation process, and record it as the buffer number. If the buffer number is higher than the preset number of monitoring time periods, then the task allocation process is given an early warning.

[0068] It should be noted that the early warning process for the task allocation process involves recording the duration and frequency of abnormal indicators in real time and issuing alerts to system administrators through a visual interface, so that timely measures can be taken to prevent the task allocation process from deteriorating further.

[0069] It should be noted that, Figure 2 The diagram illustrates the logical flow of the intelligent psychological state assessment system for students. First, multimodal data is collected in the AI ​​tree hole environment, and the quality indicators are dynamically repaired with the help of the data analysis and correction module to ensure the reliability of the subsequent calculation basis. The cluster analysis module processes the corrected data into clusters, and then accurately identifies similar behaviors or expression patterns, which significantly improves the focus of model recognition and the rationality of intervention resource allocation. Subsequently, by accurately executing task allocation, the system avoids blind resource occupation, and promptly performs load transfer and node thermal control adjustment when local congestion, abnormal hot spots or resource bottlenecks are detected.

[0070] It should be noted that, Figure 3 This is a schematic diagram of the logical flow of the dynamic adjustment of student publication data processing resources according to the present invention. First, the system extracts the quality compliance characterization value of student publication data and compares it with a set threshold. If it is higher than or equal to the threshold, a first correction is performed; otherwise, a second correction is performed. The corrected data undergoes cluster analysis to obtain cluster similarity confidence, and based on this, student publication data sequence clusters are selected, including a first data sequence cluster and a second data sequence cluster. For the first data sequence cluster, the system dynamically adjusts the real-time analysis frequency of its data processing resources; for the second data sequence cluster, the system dynamically adjusts its data sampling rate. By introducing a comparison mechanism between the data quality characterization value and the threshold, dynamic data correction is achieved; by obtaining cluster similarity confidence through cluster analysis, valuable data sequence clusters are accurately selected; and by adjusting the real-time analysis frequency and data sampling rate of data processing resources for different types of data sequence clusters, the system resource allocation and utilization efficiency are optimized.

[0071] 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 a completely hardware embodiment, a completely software embodiment, or an embodiment 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, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0072] 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.

[0073] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0074] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A student intelligent psychological state assessment system, characterized in that, include: The data analysis and correction module is used to obtain the student posting data in the AI ​​Tree Hole for preliminary analysis, obtain the student posting data quality compliance characterization value, and correct the student posting data based on the student posting data quality compliance characterization value; The specific process of correcting student-published data based on the quality conformity characterization value is as follows: A preset monitoring period is set. During the monitoring period, the student's published data quality compliance value is extracted and compared with the set student published data quality compliance threshold. If the student's published data quality compliance value is higher than or equal to the student published data quality compliance threshold, a first correction is performed. If the student's published data quality compliance value is lower than the student published data quality compliance threshold, a second correction is performed. The cluster analysis module is used to extract corrected student publication data and perform cluster analysis to obtain the cluster similarity confidence of student publication data. Based on the cluster similarity confidence of student publication data, the data is filtered to obtain student publication data sequence clusters, and the processing resources for student publication data are dynamically adjusted. The process of filtering student publication data based on cluster similarity confidence to obtain student publication data sequence clusters is as follows: The corrected student publication data is extracted and cluster analysis is performed to obtain the cluster similarity confidence of the student publication data. Based on the cluster similarity confidence of the student publication data, the student publication data sequence cluster is obtained, which includes the student's first publication data sequence cluster and the student's second publication data sequence cluster. The task allocation and scheduling module is used to analyze the scheduling data of the student's published data sequence clusters to obtain the scheduling exposure index of the student's published data sequence clusters, and to allocate and schedule tasks for the student's published data based on the scheduling exposure index of the student's published data sequence clusters. The task allocation early warning module is used to acquire and analyze key indicator data in the task allocation process, and combine it with the scheduling exposure index of the student's published data sequence cluster to obtain the task allocation process abnormality index. Based on the task allocation process abnormality index, screening is performed, and early warnings are issued for the task allocation process.

2. The student intelligent psychological state assessment system as described in claim 1, characterized in that: The specific process for obtaining the quality conformity value of student-published data is as follows: The data published by students in the AI ​​Tree Hole is extracted, including the peak signal-to-noise ratio (PSNR) of images, contrast ratio, sentence structure completeness rate, and number of redundant reduplicated words. The proportions of the peak signal-to-noise ratio of images with the defined peak signal-to-noise ratio, contrast ratio with the defined contrast ratio, sentence structure completeness rate with the defined sentence structure completeness rate, and number of redundant reduplicated words with the defined redundant reduplicated words are analyzed and weighted by coefficients to obtain the quality compliance characterization value of the student published data. The quality compliance characterization value of the student published data is used to evaluate the information reliability of the student published data.

3. The student intelligent psychological state assessment system as described in claim 1, characterized in that: The specific process for dynamically adjusting the data processing resources for student publications is as follows: Extract the first and second data sequence clusters of student publications. Dynamically adjust the real-time analysis frequency of the student publication data processing resources in the first data sequence cluster and the data sampling rate of the student publication data processing resources in the second data sequence cluster.

4. The student intelligent psychological state assessment system as described in claim 1, characterized in that: The specific process for obtaining the scheduling exposure index of the student's published data sequence cluster is as follows: During the monitoring period, the cumulative server throughput, cumulative memory dwell time, and cumulative propagation times corresponding to the student's published data are obtained. The cumulative server throughput, cumulative memory dwell time, and cumulative propagation times corresponding to the student's published data are compared with the cumulative throughput of server resources, the cumulative memory dwell time, and the cumulative propagation times. After introducing weight coefficients, these are coupled to obtain the scheduling exposure index of the student's published data sequence cluster. The scheduling exposure index of the student's published data sequence cluster is used to evaluate the degree of system resource occupation of the data sequence cluster.

5. The student intelligent psychological state assessment system as described in claim 1, characterized in that: The specific process of allocating and scheduling student publication data based on the scheduling exposure index of the student's publication data sequence cluster is as follows: Extract the scheduling exposure index of the student's published data sequence cluster and compare it with the set scheduling exposure index threshold of the student's published data sequence cluster. If the scheduling exposure index of the student's published data sequence cluster is higher than or equal to the scheduling exposure index threshold of the student's published data sequence cluster, the content corresponding to the scheduling exposure index of the student's published data sequence cluster will be preferentially allocated to the real-time processing queue.

6. The student intelligent psychological state assessment system as described in claim 1, characterized in that: The specific process for obtaining the task allocation process anomaly index is as follows: Key indicator data in the task allocation process are extracted during the monitoring period, including task allocation success rate, task allocation response time, and task completion rate. The task allocation success rate is compared with the defined task allocation success rate, the task allocation response time is compared with the defined task allocation response time, and the task completion rate is compared with the defined task completion rate. The scheduling exposure index and weight coefficient of the student's published data sequence cluster are introduced and coupled to obtain the task allocation process anomaly index. The task allocation process anomaly index is used to evaluate the stability of the task allocation mechanism.

7. The student intelligent psychological state assessment system as described in claim 1, characterized in that: The screening process based on the anomaly index of the task allocation process is as follows: A preset monitoring period is set, and the task allocation process abnormality index is obtained for each monitoring time period within the monitoring period. The number of monitoring time periods in which the task allocation process abnormality index is higher than or equal to the task allocation process abnormality index threshold is counted.

8. The student intelligent psychological state assessment system as described in claim 7, characterized in that: The process of issuing early warnings for the task allocation process is as follows: Extract the number of monitoring time periods in each monitoring time period where the abnormal index of the task allocation process is higher than or equal to the threshold of the abnormal index of the task allocation process, and record it as the buffer number. If the buffer number is higher than the preset number of monitoring time periods, then issue an early warning for the task allocation process.

Images