University performance evaluation management system and method based on multi-source data fusion

Abstract

The application discloses a college performance evaluation management system and method based on multi-source data fusion, and relates to the technical field of college performance evaluation management.The system comprises a target setting and disassembly module, a disassembly rationality evaluation module, a target progress early warning module and a performance accounting and improvement module.The application assigns initial indexes to corresponding executors, then performs weight distribution on the initial indexes of each target management level to obtain corresponding initial weights, performs multi-level index disassembly rationality evaluation, judges whether to perform index disassembly optimization, then performs performance accounting, judges whether to push early warning information to a preset person in charge, finally performs performance accounting update evaluation after performance accounting, and judges whether there is a demand to generate a performance accounting update scheme, thereby improving the effectiveness of college performance evaluation management and solving the problem of low effectiveness of college performance evaluation management in the prior art.

Description

Technical Field

[0001] This invention relates to the field of university performance evaluation management technology, and in particular to a university performance evaluation management system and method based on multi-source data fusion. Background Technology

[0002] The purpose of performance evaluation management in higher education institutions is to comprehensively evaluate teachers and administrators through a scientific and reasonable assessment system, thereby stimulating the work enthusiasm and creativity of all members within the institution. The work undertaken by university teachers has become more diverse and complex, involving multiple areas such as teaching, research, and service. This diversification of work nature makes it difficult for evaluation data from a single source to comprehensively and objectively reflect teachers' actual work performance. Therefore, multi-source data evaluation can collect information from different angles and levels, including multiple dimensions such as teaching competitions and academic papers, thus forming a more comprehensive and objective evaluation result. This method can avoid the subjectivity and bias that may arise from a single evaluation subject, improving the fairness and credibility of the evaluation.

[0003] Combining the Objectives and Key Results (OKR) methodology and the Plan-Do-Check-Act (PDCA) framework, a comprehensive and systematic management mechanism is constructed through four core processes: objective decomposition, objective execution, dynamic monitoring, and performance evaluation and feedback. OKR aligns objectives and key results at each level by setting clear goals and quantifiable key outcomes, and supports full organizational transparency by allowing all employees to view the results. The PDCA closed-loop mechanism is manifested in the following stages: clarifying objectives, key outcomes, timelines, and responsible parties in the objective decomposition and setting phase; recording progress through data-driven and manual reporting in the objective execution phase; calculating progress in real time, triggering alerts, and displaying results through dashboards in the dynamic monitoring phase; and analyzing problems based on performance evaluation results, developing improvement measures, and providing feedback for the next round of objective setting in the performance evaluation and feedback phase. Continuous optimization is achieved through the continuous cycle of these four stages.

[0004] For example, Chinese patent application CN116307841A discloses a method for constructing a multi-indicator evaluation system for universities, which includes: Step 1, obtaining publicly available data; Step 2, establishing an evaluation indicator system, cleaning and standardizing the data within the indicator system; proposing a multi-indicator evaluation method, and calculating the weight of each indicator in the indicator system using this method; calculating the performance score of the collaborative innovation center using an ideal solution method; Step 3, combining the method in Step 2 and integrating various typical multi-indicator evaluation methods to construct a performance evaluation system for the collaborative innovation center; Step 4, dynamically updating the performance evaluation system for the collaborative innovation center based on an optimal selection strategy, and dynamically recommending evaluation methods according to user needs and concerns; Step 5, constructing a visualization system for the performance evaluation of university collaborative innovation.

[0005] For example, Chinese patent application CN119005816B discloses a performance evaluation system and method, which includes: creating a dataset of individual performance evaluation indicators and assigning weights to the individual performance evaluation indicators; assigning values ​​to each individual performance evaluation indicator in order of their ranking and calculating the total performance evaluation value for each person; dividing all participants into groups based on the projects they participated in, extracting the project-related performance evaluation values ​​of project participants based on the projects, and assigning corresponding weights to the performance evaluation values ​​of the relevant individual performance evaluation indicators; calculating the performance evaluation value for each project; extracting the corresponding project performance evaluation values ​​for each college and calculating the performance evaluation value for each college; assigning weights to the performance evaluation values ​​of each college and calculating the performance evaluation value for each college.

[0006] The above-mentioned technology has at least the following technical problems:

[0007] In existing technologies, during multi-level decomposition, the weights corresponding to indicators in the first level are passed down level by level. This can cause the weights of the same indicator in subsequent levels to be amplified or reduced, potentially resulting in subsequent weights being too high or too low, failing to reflect the actual contribution. This is especially problematic when multiple colleges share the same indicator, where the resulting weights may be greater than or less than 100%.

[0008] Furthermore, as OKR implementation progresses, the number and complexity of indicators to be calculated at each level increase rapidly. Due to the diversity of colleges and the complexity of objectives in universities, problems such as inconsistent target cycles and dynamic adjustments to indicator weights arise in university performance evaluation management. For example, the objectives, implementation cycles, and feedback mechanisms of each college vary greatly. Research colleges may adjust their objectives quarterly, while teaching colleges may adjust them semester by semester. This causes the weight of teaching and research to change with the objectives of different semesters and years. Each adjustment affects the performance calculation results. Overly frequent adjustments not only affect the stability of the data but also lead to an overly reactive feedback mechanism and may affect the continuity of historical trends, resulting in low effectiveness of university performance evaluation management. Summary of the Invention

[0009] To address the technical problem of low effectiveness in university performance evaluation management in existing technologies, this invention provides a university performance evaluation management system and method based on multi-source data fusion. The technical solution is as follows:

[0010] On the one hand, a university performance evaluation and management system based on multi-source data fusion is provided. This system includes: a target setting and decomposition module, a decomposition rationality assessment module, a target progress early warning module, and a performance accounting and improvement module. The target setting and decomposition module is used to obtain initial indicators reflecting the completion of teaching tasks, faculty research activities, and teaching quality feedback for a designated university. These initial indicators are then allocated to the corresponding implementing parties according to the target management hierarchy, which includes the university level, college level, and faculty level. The decomposition rationality assessment module is used to assign weights to the initial indicators at each target management level to obtain corresponding initial weights reflecting the contribution of each initial indicator to the annual target, and to conduct a multi-level indicator decomposition rationality assessment to measure... The rationality of the initial indicator decomposition and weight allocation process is assessed. Based on the rationality evaluation results, it is determined whether there is a need for indicator decomposition optimization. If so, performance accounting is performed based on the optimized indicators used to update performance accounting and the corresponding optimized weights. Otherwise, performance accounting is performed based on the initial indicators and the corresponding initial weights. The target progress early warning module is used to determine whether to push early warning information to the preset responsible persons based on the performance accounting results and the indicator deviation coefficient used to reflect the degree of deviation between each indicator and the corresponding reference indicator. The performance accounting and improvement module is used to perform performance accounting update evaluation based on the parameters of the performance accounting process and to determine whether there is a need to generate a performance accounting update plan to improve the accuracy of performance accounting.

[0011] On the other hand, a performance evaluation and management method for universities based on multi-source data fusion is provided. This method includes: S1, obtaining initial indicators reflecting the completion of teaching tasks, teachers' research activities, and teaching quality feedback of a designated university, and allocating the initial indicators to the corresponding implementers according to the target management level, which includes the university level, college level, and teacher level; S2, assigning weights to the initial indicators at each target management level to obtain corresponding initial weights to reflect the contribution of each initial indicator to the annual target, conducting a multi-level indicator decomposition rationality assessment to measure the rationality of the initial indicator decomposition and weight allocation process, and judging whether there is a need for indicator decomposition optimization based on the obtained rationality assessment results. If so, performance accounting is carried out according to the optimized indicators used to update performance accounting and the corresponding optimized weights; otherwise, performance accounting is carried out according to the initial indicators and the corresponding initial weights; S3, judging whether to push early warning information to the preset responsible persons based on the performance accounting results and the indicator deviation coefficient used to reflect the degree of deviation between each indicator and the corresponding reference indicator; S4, conducting a performance accounting update assessment based on the parameters of the performance accounting process, and judging whether there is a need to generate a performance accounting update plan to improve the accuracy of performance accounting.

[0012] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0013] 1. In the digital campus phase, universities formulate annual goals and break them down to various departments and faculty members. Due to the different implementing entities at each level, the weights cannot accurately reflect the actual contributions of each faculty member. Obtaining initial indicators ensures their comprehensiveness and accuracy, accurately reflecting the university's operational status in all aspects. Allocating indicators according to the goal management hierarchy clarifies the responsibilities and tasks of each implementing party, making goal management clearer and more orderly, thus improving the efficiency and effectiveness of goal achievement. Then, the initial indicators are weighted, and the rationality of multi-level indicator decomposition is evaluated to determine whether further indicator decomposition optimization is needed, enabling a scientific approach. By accurately measuring the contribution of each indicator to the annual goals, problems in the indicator breakdown and weight allocation process can be identified and resolved in a timely manner, thereby improving the accuracy and reliability of performance accounting. Then, based on the performance accounting results and indicator deviation coefficients, early warning judgments can be made to promptly identify deviations and problems in the goal execution process and push early warning information to the responsible persons in a timely manner to prevent the problems from escalating further, thereby improving the timeliness and effectiveness of university performance evaluation management. Finally, through performance accounting update evaluation, shortcomings and problems in the performance accounting process can be identified in a timely manner, and update plans can be generated and implemented based on the evaluation results, which can continuously improve the accuracy of performance accounting.

[0014] 2. By comparing the reasonable indices of the indicator decomposition, it is possible to quickly determine whether the indicator decomposition needs optimization, avoiding unnecessary optimization processes and improving the efficiency of university performance evaluation management. Then, using directed graph modeling to model indicator dependencies clearly represents the complex relationships between indicators. The depth-first search algorithm can efficiently detect circular dependencies, avoiding logical errors in indicator decomposition and ensuring the rationality and executability of the indicator system. Finally, by checking the total weight and performing normalization, the rationality of weight allocation is ensured, making each indicator comparable in the overall performance evaluation. Due to the uneven amount of data collected for each indicator, the importance and weight of some indicators are mismatched. Therefore, sorting the indicators according to priority highlights important indicators. By setting a minimum weight for core indicators, it is ensured that core indicators receive sufficient attention in weight allocation, avoiding the problem of mismatch between importance and weight caused by uneven data volume. Furthermore, since certain core indicators are distributed across multiple colleges or individuals, it is necessary to ensure the fairness and rationality of weight allocation. Therefore, splitting the weights according to the contribution ratio of each college reflects the actual contribution of each college in completing the indicators, thus ensuring the fairness and rationality of weight allocation. Normalizing the split weights further ensures the accuracy of weight allocation. In the process of multi-level indicator decomposition and weight allocation, if the weight of the underlying indicators is too low, it may be almost meaningless in actual implementation. Therefore, by setting a minimum weight threshold, the situation where the weight of the underlying indicators is too low and becomes meaningless in actual implementation is avoided. Allocating the remaining weights according to the weight ratio ensures the rationality and fairness of weight adjustment.

[0015] 3. By comparing updated performance evaluation indicators, a basis for subsequent decision-making is provided. Then, based on different thresholds and update cycles, it is determined whether to generate and send updated performance evaluation plans. This ensures timely response to changes in performance evaluation indicators, allowing managers to understand the situation and take action immediately, improving the timeliness of performance management. Finally, when the average target achievement rate is less than the minimum threshold, managers are prompted to identify problems and improve the plan in a timely manner. When the average target achievement rate is greater than the maximum threshold, the performance evaluation plan is optimized to make the weight allocation more reasonable and more accurately reflect the contribution of each indicator to performance, improving the fairness of performance evaluation. When the average target achievement rate meets the target, no updated performance evaluation plan is generated, reducing unnecessary plan updates, lowering management costs, and allowing managers to understand the current performance status, providing a reference for subsequent decision-making. When the standard deviation of the target achievement rate is greater than the standard deviation of the reference target achievement rate, or the target correlation coefficient is less than the reference target correlation coefficient, a performance evaluation update plan cannot be generated based on the target achievement rate. The weights need to be adjusted using the target correlation coefficient to make the performance evaluation more reflective of the importance of indicators to the annual target. By mapping to obtain the calculation time window, the data range of the correlation analysis is ensured to be reasonable, improving the accuracy of university performance evaluation. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is one of the flowcharts for university performance evaluation management provided in this embodiment of the invention;

[0018] Figure 2 This is the second flowchart of the university performance evaluation management provided in this embodiment of the invention;

[0019] Figure 3 This is a schematic diagram of the structure of the university performance evaluation management system based on multi-source data fusion provided in an embodiment of the present invention;

[0020] Figure 4 This is an overall architecture diagram of the university performance management system provided in this embodiment of the invention;

[0021] Figure 5 This is an interface diagram of performance reporting in the university performance management system provided in this embodiment of the invention;

[0022] Figure 6 This is an interface diagram of the teaching reform project reporting dialog box in the university performance management system provided in this embodiment of the invention;

[0023] Figure 7 This is an interface diagram of the dialog box for issuing annual performance appraisals in the university performance management system provided in this embodiment of the invention;

[0024] Figure 8 This is an interface diagram of teaching performance calculation in the university performance management system provided in this embodiment of the invention;

[0025] Figure 9 The flowchart of the university performance evaluation and management method based on multi-source data fusion provided in this embodiment of the invention. Detailed Implementation

[0026] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0027] Example 1: As Figure 1 One of the flowcharts for university performance evaluation management is shown below, such as Figure 2The second flowchart of the university performance evaluation management process illustrates the following logic: Initial indicators at each target management level are weighted to obtain corresponding initial weights. A multi-level indicator decomposition rationality assessment is then conducted. If the indicator decomposition rationality index is greater than the indicator decomposition rationality threshold, performance calculation is performed based on the initial indicators and corresponding initial weights. Otherwise, indicator decomposition optimization is performed, and performance calculation is performed based on the optimized indicators and corresponding optimized weights. The specific process for indicator decomposition optimization is as follows: Initial indicators at each level are modeled as directed graphs. It is determined whether circular dependencies exist. If circular dependencies exist, saving is prohibited, and preset personnel are prompted to modify the dependency relationships. Otherwise, if the current level... If the sum of the weights for each level is not 1, then each weight is normalized; otherwise, the preset personnel are prompted to manually modify the initial weights. Then, if the performance score is less than the first reference performance score, a warning message is directly sent to the preset responsible person. If the performance score is less than the second reference performance score, and the number of times it exceeds the second reference performance score is greater than a preset number, a warning message is directly sent to the preset responsible person. Otherwise, if there is an initial indicator with a deviation coefficient greater than the first indicator deviation threshold, a warning message is directly sent to the preset responsible person; if there is an initial indicator with a deviation coefficient greater than the second indicator deviation threshold, and the number of times it exceeds the optimized indicator deviation threshold is greater than a preset number, a warning message is directly sent to the preset responsible person. Within a preset cooling-off period after the warning information is pushed, the warning information will not be pushed to the preset responsible person. Finally, a performance accounting update evaluation will be conducted. If the performance accounting update indicator is greater than the first performance accounting update threshold, a performance accounting update plan will be generated and sent to the preset performance management personnel. If the performance accounting update indicator is greater than the second performance accounting update threshold, and the performance accounting update period is greater than the preset performance accounting update period, a performance accounting update plan will be generated and sent to the preset performance management personnel. Otherwise, performance accounting will continue according to the current performance accounting update plan, and it will be determined whether a rejection notice for the performance accounting update plan has been received from the preset performance management personnel. If so, continue performance accounting according to the current performance accounting version; otherwise, update the performance accounting version according to the generated performance accounting update plan, map historical indicators, retain historical versions, and clear the cache of historical indicators whose retention time is longer than the preset retention time. The specific process for generating the performance accounting update plan is as follows: if the average target achievement rate is less than the minimum threshold of the target achievement rate, directly prompt the preset performance management personnel to update the performance accounting plan; if the average target achievement rate is greater than the maximum threshold of the target achievement rate, map to obtain the weight correction factor, and adjust the initial weight according to the weight correction factor; otherwise, do not generate a performance accounting update plan and provide feedback to the preset performance management personnel.

[0028] like Figure 3The diagram shown illustrates the structure of a university performance evaluation management system based on multi-source data fusion. This embodiment of the invention provides a university performance evaluation management system based on multi-source data fusion, which includes: a target setting and decomposition module, a decomposition rationality assessment module, a target progress early warning module, and a performance accounting and improvement module.

[0029] like Figure 4 The overall architecture diagram of the university performance management system shown is as follows: It adopts the concept of "front-end and back-end separation" and fully considers the ease of use and concurrency of the system. It provides two sets of applications, one for computers and one for mobile devices, to meet different usage scenarios. It integrates four major modules: personnel management, goal management, scientific research management and performance management. It automatically captures personnel, academic affairs and scientific research data, and combines them with manual data entry to build a complete data pool. Information exchange is achieved through data interfaces, and a microservice architecture (PostgreSQL database + message middleware) is adopted to support high-concurrency data processing.

[0030] The goal setting and decomposition module combines Objectives and Key Results (OKR) and Plan-Do-Check-Act (PDCA) to break down annual goals into initial indicators. This allows for the acquisition of initial indicators that reflect the completion of teaching tasks, faculty research activities, and teaching quality feedback for a designated university. For example, "improving undergraduate teaching quality" can be broken down into "average course evaluation score," "course excellence rate," and "faculty teaching training coverage rate." The initial indicators are then allocated to the corresponding implementers according to the goal management hierarchy to clarify the specific tasks of each level in achieving the overall goals of the university. The goal management hierarchy includes the university level, the college level, and the faculty level.

[0031] See also Figure 5 The interface diagram shown is of the performance reporting system in the university performance management system, and as shown below... Figure 6 The screenshot shows the interface of the teaching reform project reporting dialog box in the university performance management system. The specific operation is as follows: On the faculty / staff side, faculty / staff log in to the system monthly and click on

Employee Points Performance

Performance Reporting

Teaching Reform Projects

Add

[0032] like Figure 7 The image shown is an interface screenshot of the annual performance appraisal distribution dialog box in the university performance management system. The specific operation is as follows: On the administrator's side, the university administrator clicks the "Distribute" button on the annual performance distribution page to distribute the performance reporting task for the current year. In the pop-up annual performance appraisal distribution dialog box, the administrator selects the assessment year and the start and end dates for the reporting, and then clicks "Submit". After successful distribution, the system will automatically load the performance reporting data distributed to each faculty and staff member.

[0033] Initial indicators were obtained by using multi-source data, ensuring their comprehensiveness and accuracy and enabling them to truly reflect the operational status of various aspects of universities. The allocation of indicators according to the target management hierarchy clarified the responsibilities and tasks of each implementing party, making target management clearer and more orderly, which is conducive to improving the efficiency of university performance evaluation management.

[0034] The decomposition rationality assessment module is used to assign weights to initial indicators at each target management level to obtain corresponding initial weights reflecting the contribution of each initial indicator to the annual target. It performs a multi-level indicator decomposition rationality assessment to measure the rationality of the initial indicator decomposition and weight allocation process. Based on the obtained rationality assessment results, it determines whether there is a need for indicator decomposition optimization. If so, performance accounting is performed based on the optimized indicators and corresponding optimized weights used to update performance accounting; otherwise, performance accounting is performed based on the initial indicators and corresponding initial weights. Weighting the initial indicators scientifically measures the contribution of each indicator to the annual target, making performance evaluation more objective and fair. The multi-level indicator decomposition rationality assessment can promptly identify problems in the indicator decomposition and weight allocation process. By optimizing indicators and weights, it improves the accuracy and reliability of performance accounting, providing more valuable references for university management decisions.

[0035] The target progress early warning module determines whether to send early warning information to designated responsible persons based on performance accounting results and indicator deviation coefficients, which reflect the degree of deviation between each indicator and its corresponding reference indicator. This alerts the responsible persons to problems arising during target execution. Early warning judgments based on performance accounting results and indicator deviation coefficients can promptly identify deviations and problems in the target execution process and send early warning information to the responsible persons in a timely manner. This helps the responsible persons take timely measures to adjust and improve, prevent problems from escalating, ensure the smooth achievement of targets, and improve the timeliness and effectiveness of university management.

[0036] The performance accounting and improvement module is used to conduct performance accounting update assessments based on parameters from the performance accounting process, and to determine whether there is a need to generate performance accounting update plans, thereby improving the accuracy of performance accounting. Through performance accounting update assessments, deficiencies and problems in the performance accounting process can be identified in a timely manner. Based on the assessment results, update plans can be generated and implemented, continuously improving the accuracy of performance accounting. This ensures that performance evaluation results more realistically and accurately reflect the actual operation of universities, providing strong support for university management optimization and sustainable development.

[0037] In this embodiment, the existing performance evaluation mechanism mainly focuses on content collection and display, lacking analysis of the alignment and support between top-down and bottom-up goals. This results in current performance evaluations exhibiting low efficiency, weak support between macro and micro goals, delayed problem identification, and untimely and inaccurate data. Consequently, it is difficult to comprehensively understand the situation of faculty and staff, let alone provide targeted guidance and improvement. The performance evaluation system is incomplete, and related data exhibits significant lag, making it impossible to promptly understand the progress of performance evaluation through various data and indicators throughout the process. Furthermore, due to the lack of process data support, unscientific performance evaluation methods, and weak performance management functions, the comprehensive application of the evaluation process and results is neglected, making it difficult to achieve real-time, refined, and scientific process control. In the digital campus stage, the methods for summarizing process and result data in performance evaluation remain relatively simple, involving a large amount of manual statistical work. This method is not only inefficient and error-prone, but more importantly, it lacks real-time data and comprehensive analytical dimensions, making it difficult to gain a comprehensive and detailed understanding through data, and even more difficult to use data to assist in relevant decision-making.

[0038] By combining OKR with PDCA through the above steps, a digital closed-loop system for university performance evaluation is constructed, achieving comprehensiveness, scientific rigor, timeliness, and accuracy in university performance evaluation. Starting with goal setting and decomposition, it ensures the reasonable allocation of goals and clear responsibilities; through the rationality assessment of decomposition, it improves the basic quality of performance accounting; the goal progress early warning module can promptly identify problems and remind responsible parties; and the performance accounting and improvement module continuously improves the accuracy of performance accounting, thereby enhancing the effectiveness of university performance evaluation management.

[0039] Furthermore, a rationality assessment of the multi-level indicator decomposition is conducted, using the following specific methods:

[0040] The multi-path dependency conflict coefficient is obtained by comparing the number of repeatedly referenced initial indicators with the total number of initial indicators. This coefficient is used to assess the complexity of dependencies among initial indicators during multi-level decomposition. A higher multi-path dependency conflict coefficient (i.e., a larger proportion of repeated references) leads to the same indicator being weighted repeatedly in multiple colleges or paths, thus increasing the level-level weight deviation coefficient. For example, if a teaching reform project is simultaneously referenced by the Academic Affairs Office, Student Affairs Office, School of Information Engineering, and School of Management, with the Academic Affairs Office having a weight of 40% and a completion rate of 90%, the Student Affairs Office having a weight of 40% and a completion rate of 85%, the School of Information Engineering having a weight of 30% and a completion rate of 80%, and the School of Management having a weight of 30% and a completion rate of 75%, without weight splitting, the total weight of the next level will exceed 100%, increasing the level-level weight deviation coefficient. This, in turn, leads to a higher overall completion rate. If it exceeds 100%, the deviation coefficient of the aggregated indicator results will increase.

[0041] The relative deviation between the total weight of the current level and the total weight of the next level is processed to obtain the level weight deviation coefficient. The level weight deviation coefficient is used to measure the degree of deviation between the current level and the next level in weight allocation. If the level weight deviation coefficient is higher, it will usually lead to a higher deviation coefficient of the index summary result, because the unbalanced weight allocation will directly affect the weighted result of the next level.

[0042] The relative deviation of the initial indicator summary results at the current level and the next level is processed to obtain the indicator summary result deviation coefficient. The initial indicator summary result represents the result obtained by weighting all initial indicators and their corresponding initial weights at the same level. The indicator summary result deviation coefficient is used to evaluate the difference between the current level and the next level in the indicator summary results.

[0043] By introducing a reasonable evaluation and balancing weight, the multi-path dependency conflict index, the hierarchical weight deviation coefficient, and the index aggregation result deviation coefficient are subjected to inverse proportional assignment coupling processing to obtain the index decomposition rationality index. The index decomposition rationality index is used to quantitatively evaluate the rationality of the initial multi-level index decomposition and weight allocation process. The specific constraint expression of the index decomposition rationality index is as follows: In the formula, F represents the index decomposition rationality index, F1 represents the multi-path dependency conflict coefficient, F2 represents the hierarchical weight deviation coefficient, F3 represents the index summary result deviation coefficient, α1 represents the first rational assessment balance weight, α2 represents the second rational assessment balance weight, and α3 represents the third rational assessment balance weight.

[0044] The reasonable assessment balance weights involved are obtained from a preset database and set by preset personnel. Specifically, they include a first reasonable assessment balance weight, a second reasonable assessment balance weight, and a third reasonable assessment balance weight, the sum of which is 1. For example, the multi-path dependency conflict coefficient and the first reasonable assessment balance weight form a corresponding first indicator decomposition reasonable mapping set. The real-time multi-path dependency conflict coefficient is input into the first indicator decomposition reasonable mapping set to obtain the first reasonable assessment balance weight. The first reasonable assessment balance weight represents the degree of influence of the multi-path dependency conflict coefficient on the indicator decomposition reasonable index. The second reasonable assessment balance weight represents the degree of influence of the hierarchical weight deviation coefficient on the indicator decomposition reasonable index. The third reasonable assessment balance weight represents the degree of influence of the indicator aggregation result deviation coefficient on the indicator decomposition reasonable index.

[0045] In this embodiment, through the above steps, the multi-level decomposition and weight allocation process of the initial indicators can be comprehensively evaluated from multiple dimensions. By reasonably evaluating and balancing the weights, problems in the indicator decomposition and weight allocation process can be identified in a timely manner, providing a basis for subsequent indicator decomposition optimization and improving the effectiveness of university performance evaluation.

[0046] Furthermore, based on the obtained rationality assessment results, it is determined whether there is a need for indicator decomposition and optimization. The specific process is as follows:

[0047] If the reasonable index of indicator decomposition is greater than the reasonable threshold of indicator decomposition, no indicator decomposition optimization will be performed; otherwise, indicator decomposition optimization will be performed. The reasonable threshold of indicator decomposition is set by preset personnel. By comparing the reasonable index of indicator decomposition, it is possible to quickly determine whether indicator decomposition optimization is needed, avoiding unnecessary optimization processes and improving the efficiency of university performance evaluation management.

[0048] Specifically, the optimization of indicator decomposition involves: modeling the initial indicators at each level as directed graphs and using depth-first search to detect the existence of circular dependencies; in university performance evaluation management, directed graphs are used to model the initial indicators at each level, with vertices representing indicators and directed edges representing the dependencies between indicators, which can clearly represent the relationships between indicators; depth-first search is used to detect whether there are circular dependencies in the directed graph, that is, starting from an indicator, can it return to itself through the dependency relationship? This can efficiently detect circular dependencies, avoid logical errors in indicator decomposition, and ensure the rationality and executability of the indicator system.

[0049] If circular dependencies exist, saving is prohibited, and the designated personnel are prompted to modify the dependencies; otherwise, initial weight optimization is performed. Initial weight optimization specifically involves: if the sum of the weights at the current level is not 1, normalizing each weight; otherwise, the designated personnel are prompted to manually modify the initial weights (i.e., there may be other unreasonable weights requiring manual intervention). Normalization is used to adjust the weights so that their sum is 1. By checking the total weights and performing normalization, the rationality of weight allocation is ensured, making each indicator comparable in the overall university performance evaluation.

[0050] In this embodiment, the above steps enable the construction of a reasonable, scientific, fair, and executable multi-level indicator system, which avoids logical errors in indicator decomposition, such as circular dependencies, thereby achieving the effectiveness and accuracy of university performance evaluation.

[0051] like Figure 8 The image shown is an interface diagram of the teaching performance calculation in the university performance management system. After the teaching performance data submitted by faculty and staff is approved by the Academic Affairs Office, the system can automatically calculate the archived teaching performance scores for each faculty and staff member, which can be viewed by the university administrator.

[0052] To further clarify, the specific process for determining whether to send an early warning message to the designated responsible party is as follows:

[0053] The performance score is obtained by assigning and coupling initial indicators (or optimized indicators; all initial indicators below are treated similarly. If indicators are decomposed and optimized, they are optimized indicators; otherwise, they are initial indicators) and their corresponding initial weights (or optimized weights; all initial weights below are treated similarly. If indicators are decomposed and optimized, they are optimized weights; otherwise, they are initial weights) to each indicator. The performance score is obtained by multiplying the initial indicators and initial weights. The performance score is used to measure the overall performance level of each indicator within a specific evaluation period. By comprehensively calculating multiple initial indicators according to their weights, the overall performance level of each indicator within a specific evaluation period can be comprehensively and objectively reflected, which is convenient for comparison and analysis.

[0054] If the performance score is lower than the first reference performance score, an early warning message will be sent directly to the designated person in charge, and the performance score of that staff member will be marked in red. The first reference performance score, the second reference performance score, the preset number of times, the first indicator deviation threshold, the second indicator deviation threshold, and the preset cooling-off period are set by the designated personnel. This allows for the rapid identification of evaluation subjects with significantly low performance levels, timely sending of early warning messages to the responsible person, prompting the responsible person to pay attention and take measures to improve performance, effectively stimulating and maintaining the work enthusiasm of staff, thereby improving teaching quality and work efficiency.

[0055] If the performance score is lower than the second reference performance score, and the number of times it exceeds the second reference performance score is greater than a preset number, an early warning message will be sent directly to the preset responsible person, and the performance score of that faculty member will be marked in red. Otherwise, the decision to send an early warning message to the preset responsible person will be based on the indicator deviation coefficient. The first reference performance score is lower than the second reference performance score. The indicator deviation coefficient represents the result obtained by performing relative deviation processing on the initial indicator and the corresponding reference indicator. The relative deviation processing in this application means that the absolute value of the difference between the actual value and the reference value is used to perform a ratio calculation with the reference value. This can quickly identify evaluation objects with significantly low performance levels, promptly send early warning messages to the responsible person, prompt the responsible person to pay attention and take measures to improve performance, effectively stimulate and maintain the work enthusiasm of faculty members, thereby improving teaching quality and work efficiency.

[0056] The process for determining whether to send an early warning message to a pre-defined responsible person based on the indicator deviation coefficient is as follows:

[0057] If there is an initial indicator whose deviation coefficient is greater than the first indicator deviation threshold, an early warning message will be sent directly to the designated person in charge, and that person's indicator will be marked in red. This allows for the precise identification of staff members whose performance on a certain indicator deviates significantly from expectations, enabling targeted early warning messages to be sent. This allows the person in charge to clearly identify the problem and focus their efforts on solving key issues.

[0058] If an initial indicator has a deviation coefficient greater than the second indicator deviation threshold, and the number of times it exceeds the optimized indicator deviation threshold is greater than the preset number, an early warning message will be sent directly to the preset responsible person, and the indicator for that faculty member will be marked in red, indicating that the first indicator deviation threshold is greater than the second indicator deviation threshold. For indicators with continuous deviations across multiple indicators, timely detection and early warning can be provided to prevent the accumulation of problems from leading to an overall decline in performance. At the same time, by limiting the number of occurrences, misjudgments caused by accidental factors are reduced.

[0059] The warning information will not be sent to the designated responsible persons within a preset cooling-off period after the warning information is sent; this avoids the interference caused to the responsible persons by frequent warning information and improves the effectiveness and practicality of the warning information.

[0060] It should be added that managers can use the data dashboard to understand the progress of various indicators in real time and identify existing problems and risks.

[0061] In this embodiment, the above steps enable comprehensive, accurate, and timely monitoring of the performance of all faculty and staff in universities. It allows for flexible and reasonable dissemination of early warning information based on different performance levels and indicator deviations, enabling designated responsible persons to promptly understand the performance status, identify problems, and take targeted measures to improve performance. Simultaneously, the preset cooling-off period avoids information overload, enhancing the practicality and operability of early warning systems for university performance evaluation management.

[0062] Furthermore, performance evaluation is updated based on parameters from the performance evaluation process. The specific methods are as follows:

[0063] The annual target coverage consistency coefficient is obtained by comparing the number of initial indicators whose coverage rate, reflecting the degree of implementation of initial indicators at the teacher level, is greater than the reference indicator coverage rate with the total number of initial indicators. The indicator coverage rate represents the result obtained by comparing the number of teachers allocated to each initial indicator with the total number of teachers. The comparison process in this application refers to ratio calculation. The reference indicator coverage rate, the total number of initial indicators, the reference performance score, and the reference correlation coefficient are set by preset personnel. The annual target coverage consistency coefficient is used to evaluate the alignment of university performance evaluation and the effectiveness of target decomposition. If the indicator coverage rate is low (the target is not fully covered), some work may not be assessed, thereby increasing the average deviation coefficient of performance score (the greater the deviation between the actual score and the target). In pursuing a higher indicator coverage rate, redundant indicators may be introduced (such as adding duplicate indicators to cover all targets), increasing indicator redundancy.

[0064] The performance score deviation coefficient is obtained by performing relative deviation processing on the performance scores obtained from the faculty and their corresponding reference performance scores. The average performance score deviation coefficient is then calculated by averaging all performance score deviation coefficients. The average performance score deviation coefficient is used to assess the stability and reliability of the performance evaluation results in universities. Higher indicator redundancy may lead to duplicate assessments, amplifying the average performance score deviation coefficient (e.g., if the same work is scored by multiple indicators, the deviation from expectations will be caused by the weighting of multiple indicators).

[0065] Each initial indicator is paired with a Pearson correlation coefficient to obtain the initial indicator correlation coefficient. Indicators with initial indicator correlation coefficients greater than the reference correlation coefficient are set as redundant indicators. The number of redundant indicators obtained by statistics is compared with the total number of initial indicators to obtain the indicator redundancy. The indicator redundancy is used to evaluate the indicator independence of university performance evaluation.

[0066] By introducing updated evaluation balancing weights, and assigning values ​​to the annual target coverage consistency coefficient, the average performance score deviation coefficient, and the indicator redundancy, the assigned annual target coverage consistency coefficient is inversely proportionally calculated and then coupled with the assigned average performance score deviation coefficient and indicator redundancy to obtain the performance accounting update indicator. This performance accounting update indicator is used to quantify the probability of updating the performance accounting scheme. The specific constraint expression for the performance accounting update indicator is as follows: In the formula, G represents the performance accounting update indicator, G1 represents the annual target coverage consistency coefficient, G2 represents the average deviation coefficient of performance score, G3 represents the indicator redundancy, β1 represents the first update assessment balance weight, β2 represents the second update assessment balance weight, and β3 represents the third update assessment balance weight.

[0067] The updated assessment balance weights involved are obtained from a preset database and set by preset personnel. Specifically, they include a first updated assessment balance weight, a second updated assessment balance weight, and a third updated assessment balance weight, the sum of which is 1. For example, the annual target coverage consistency coefficient and the first updated assessment balance weight form a corresponding first updated assessment balance mapping set. The real-time annual target coverage consistency coefficient is input into the first updated assessment balance mapping set to obtain the first updated assessment balance weight. The first updated assessment balance weight represents the degree of influence of the annual target coverage consistency coefficient on the performance accounting updated indicators; the second updated assessment balance weight represents the degree of influence of the average deviation coefficient of performance scores on the performance accounting updated indicators; and the third updated assessment balance weight represents the degree of influence of indicator redundancy on the performance accounting updated indicators.

[0068] In this embodiment, the above steps can comprehensively assess the probability of updating the performance accounting scheme, providing an effective basis for generating decisions on the needs of updating the performance accounting scheme and greatly reducing the arbitrariness of decision-making. By introducing balanced weights, the focus of the evaluation is dynamically adjusted to adapt to different scenario needs, thereby improving the scientific nature and credibility of university performance evaluation management.

[0069] Furthermore, determine whether there is a need to generate a performance accounting update plan:

[0070] If the updated performance evaluation indicator exceeds the first performance evaluation update threshold, a performance evaluation update plan is generated and sent to the pre-set performance management personnel. The first performance evaluation update threshold, the second performance evaluation update threshold, the pre-set performance evaluation update cycle, the pre-set retention time, the minimum threshold for target achievement rate, and the maximum threshold for target achievement rate are set by the pre-set personnel. By comparing the updated performance evaluation indicators, a basis is provided for subsequent decisions on generating performance evaluation update plans, avoiding blind updates and improving the resource utilization efficiency of university performance evaluation.

[0071] If the updated performance accounting indicator exceeds the second performance accounting update threshold and the performance accounting update cycle exceeds the preset performance accounting update cycle, a performance accounting update plan is generated and sent to the preset performance management personnel. Otherwise, performance accounting continues to be performed according to the current performance accounting update plan. The performance accounting update cycle refers to the time interval between two performance accounting updates. This balances the flexibility and stability of performance accounting, avoids the chaos in university performance evaluation management caused by frequent updates, and ensures that the plan can be adjusted in a timely manner when necessary.

[0072] The system determines whether a rejection notification from the preset performance management personnel regarding the performance accounting update plan has been received. If so, performance accounting continues according to the current performance accounting version. Otherwise, the performance accounting version is updated according to the generated performance accounting update plan, and historical indicators are mapped to match the new performance accounting version. At the same time, historical versions are retained, and cached historical indicators with a retention time greater than the preset retention time are cleared. This respects the decisions of management personnel, ensures that the performance accounting plan meets the overall management needs of the university, and reasonably processes historical data to maintain data integrity and availability.

[0073] The performance accounting update plan is generated as follows:

[0074] If the average target achievement rate is less than the minimum target achievement rate threshold, the preset performance management personnel will be directly prompted to update the performance accounting scheme. The average target achievement rate represents the average percentage of all indicators that have achieved their targets, thereby prompting managers to identify problems and improve the scheme in a timely manner.

[0075] If the average target achievement rate is greater than the maximum target achievement rate threshold, the performance accounting update indicators and the performance input weight correction factor mapping set are mapped to obtain the weight correction factor. The weight correction factor mapping set is a set obtained from the preset database that represents the mapping relationship between the performance accounting update indicators and performance scores and the weight correction factor. The initial weights are adjusted according to the weight correction factor, that is, the adjusted weights are obtained by multiplying the initial weights and the weight correction factor, which prompts managers to identify problems and improve the plan in a timely manner; otherwise, the performance accounting update plan is not generated, and feedback is given to the preset performance management personnel, thereby prompting managers to identify problems and improve the plan in a timely manner.

[0076] In this embodiment, the above steps enable dynamic adjustment and optimization of the performance accounting scheme. It can generate updated schemes in a timely and accurate manner based on changes in performance accounting indicators and target achievement, and ensure that managers can effectively participate in the decision-making process. At the same time, the reasonable processing and maintenance of historical data ensures the continuity and availability of data, providing strong support for the performance evaluation management and decision-making of universities.

[0077] This invention provides a method for university performance evaluation and management based on multi-source data fusion. For example... Figure 9 The flowchart shown is for a university performance evaluation management method based on multi-source data fusion. This method includes:

[0078] S1 obtains initial indicators reflecting the completion of teaching tasks, faculty research activities, and teaching quality feedback of designated universities. These initial indicators are then allocated to the corresponding implementing parties according to the target management hierarchy, which includes the university level, college level, and faculty level. By obtaining multiple initial indicators, the university's situation in teaching, research, and teaching quality can be comprehensively reflected, providing rich data support for performance evaluation. Allocating initial indicators according to the target management hierarchy ensures clear goals and responsibilities for the university, colleges, and individuals, helping to form a consistent and clearly defined management system, and improving work efficiency and execution.

[0079] S2 assigns weights to the initial indicators at each target management level to reflect their contribution to the annual target. A multi-level indicator decomposition rationality assessment is conducted to evaluate the rationality of the initial indicator decomposition and weight allocation process. Based on the rationality assessment results, it is determined whether there is a need for indicator decomposition optimization. If so, performance accounting is performed based on the optimized indicators and corresponding optimized weights used to update performance accounting; otherwise, performance accounting is performed based on the initial indicators and corresponding initial weights. Weighting the initial indicators accurately reflects their contribution to the annual target, making performance evaluation more scientific and reasonable, highlighting key indicators and critical tasks. Through multi-level indicator decomposition rationality assessment and optimization, it is ensured that the indicator decomposition and weight allocation process conforms to the actual situation and performance evaluation requirements, improving the accuracy and reliability of performance accounting.

[0080] S3 determines whether to send an early warning message to the pre-assigned responsible person based on the performance accounting results and the indicator deviation coefficient used to reflect the degree of deviation between each indicator and the corresponding reference indicator; sending early warning messages based on the performance accounting results and indicator deviation coefficient can promptly identify problems and deviations in the performance completion process, remind the responsible person to take measures to adjust and improve, avoid further expansion of problems, and ensure the smooth achievement of performance goals.

[0081] S4, based on the parameters of the performance accounting process, conducts performance accounting update assessment and determines whether there is a need to generate a performance accounting update plan in order to improve the accuracy of performance accounting; through performance accounting update assessment and generation of update plans, the performance accounting process can be continuously optimized, the accuracy and scientific nature of performance accounting can be improved, and performance evaluation can better adapt to the changes and needs of university development.

[0082] Example 2: In Example 1, the importance priority of indicators was not considered during weight allocation. Due to the uneven amount of data collected for some indicators, the importance of some indicators and the weights allocated by the system are mismatched. This may result in the core indicators having too low a weight and the non-core indicators having too high a weight, failing to effectively reflect their contribution to the overall goal. The indicator decomposition and optimization process in Example 1 can be adjusted to the following process: Indicator decomposition and optimization, specifically: sort all initial indicators in descending order according to the set priority reflecting the importance of the initial indicators to the annual goal; the priority, reference priority, and core minimum weight of the initial indicators are set by preset personnel.

[0083] If an initial indicator has a higher priority than the reference priority and its corresponding initial weight is less than the set core minimum weight, the system will prompt the designated personnel to manually modify the initial weight; otherwise, each weight will be normalized. By prioritizing the initial indicators and issuing warnings for indicators that do not meet the minimum core weight, the scientific nature of the weights can be ensured, making data collection and indicator decomposition more realistic. By setting the core minimum weight, it is ensured that core indicators receive sufficient attention in weight allocation, avoiding the problem of mismatch between importance and weight due to uneven data volume.

[0084] Furthermore, since certain core indicators are distributed across multiple colleges or individuals, it is necessary to ensure the fairness and rationality of the weight allocation. The indicator decomposition optimization process in Example 1 can be adjusted to the following process:

[0085] The initial weights corresponding to each initial indicator are then distributed according to the pre-defined contribution ratio of each college (i.e., the pre-set relative proportion of each college's contribution in completing the indicator). For example, if the initial weight is 0.5, the contribution ratio of the Computer Science College is 0.2, and the contribution ratio of the Electrical Engineering College is 0.8, then the weight of the Computer Science College is... The weight of the School of Electrical Engineering is If the sum of the initial indicator's weights across all colleges after splitting is not 1, then the initial indicator's weights in each college are normalized to obtain optimized weights; otherwise, the split weights are directly set as optimized weights. Splitting weights according to the colleges' contribution proportions reflects each college's actual contribution in completing the indicator, ensuring the fairness and rationality of weight allocation. Normalizing the split weights further ensures the accuracy of weight allocation.

[0086] Example 3: In the multi-level indicator decomposition and weight allocation process of Example 1, if the weight of the bottom-level indicator may be too low, it will be almost meaningless in actual implementation and cannot reflect its actual contribution. The following process can be added to the indicator decomposition optimization in Examples 1 and 2: Indicator decomposition optimization also includes:

[0087] The initial weights of each initial indicator are sorted in descending order to obtain the minimum initial weight. The indicator with the minimum initial weight is then set as the indicator to be judged, and the remaining initial indicators are set as indicators to be assigned. If the minimum initial weight is less than the minimum weight threshold (set by a pre-defined group), the weight of the indicator to be judged is set to the minimum weight threshold, and the remaining weights are allocated to the indicators to be assigned according to the weight ratio of each initial indicator. For example, if the minimum initial weight is 0.05, the minimum weight threshold is 0.1, the initial weight of the teaching reform project score is 0.65, and the initial weight of the textbook publication score is 0.3, then the weight of the indicator to be judged is set to 0.1, and the weight of the teaching reform project score is... The weight of the textbook publication score is 1. Otherwise, no weight allocation will be performed. By setting a minimum weight threshold, the situation where the weight of the underlying indicators is too low and becomes meaningless in actual implementation is avoided; the remaining weights are allocated according to the weight ratio, ensuring the rationality and fairness of the weight adjustment, making the weight allocation of each indicator more scientific, enhancing the influence of the underlying indicators, and ensuring that even the underlying indicators with lower weights can occupy a certain weight in the overall weight system, thereby ensuring that the underlying indicators are not ignored in performance accounting.

[0088] Example 4: In Example 1, the weight allocation did not consider the correlation between indicators and target achievement rates, which may lead to excessively high weights for low-correlation indicators. Furthermore, performance evaluation updates may rely on fixed weight allocations, lacking a dynamic adjustment mechanism and failing to optimize weights based on actual performance. This lack of flexibility fails to accurately reflect the actual correlation and dynamic changes between targets. When the standard deviation of the target achievement rate is greater than the standard deviation of the reference target achievement rate, or the target correlation coefficient is less than the reference target correlation coefficient, the process for generating the performance evaluation update scheme in Example 1 can be replaced:

[0089] The performance evaluation update plan is generated as follows: Based on the calculation time window, each initial indicator and annual target is input into Pearson correlation analysis to obtain the target correlation coefficients. The current weights are then adjusted according to the output target correlation coefficients. For example, if the target correlation coefficient for the teaching reform project score is 0.35, the correlation coefficient for the textbook publication score is 0.64, and the correlation coefficient for the textbook award score is 0.87, then the weight of the teaching reform project score is... The weight of the textbook publication score is 1. The weight of the textbook award score is: The calculation time window is obtained by mapping the performance accounting update indicators and performance scores into a time window mapping set. This time window mapping set is a collection obtained from a pre-set database representing the mapping relationship between the performance accounting update indicators and performance scores and the calculation time window. By introducing Pearson correlation analysis, weights can be dynamically adjusted based on the correlation coefficient between objectives, making weight allocation more scientific and effectively reflecting the dependencies between different objectives. This allows performance accounting to more accurately reflect the relationship between indicators and annual targets. Adjusting indicator weights based on the correlation between objectives avoids over-reliance on a single objective, improving the accuracy of performance evaluation. By combining the calculation time window with timely adjustments to objective relevance, data analysis and target evaluation become more timely and reasonable, avoiding the impact of data lag or obsolescence.

[0090] 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.

Claims

1. A university performance evaluation management system based on multi-source data fusion, characterized in that: The system includes: a target setting and decomposition module, a decomposition rationality assessment module, a target progress early warning module, and a performance accounting and improvement module; The target setting and decomposition module is used to obtain initial indicators reflecting the completion of teaching tasks, teachers' research activities and teaching quality feedback of designated universities, and to allocate the initial indicators to the corresponding implementers according to the target management hierarchy, which includes the university level, college level and teacher level. The decomposition rationality assessment module is used to assign weights to the initial indicators of each target management level to obtain corresponding initial weights to reflect the contribution of each initial indicator to the annual target. It performs multi-level indicator decomposition rationality assessment to measure the rationality of the initial indicator decomposition and weight allocation process. Based on the obtained rationality assessment results, it determines whether there is a need for indicator decomposition optimization. If so, it performs performance accounting based on the optimized indicators used to update performance accounting and the corresponding optimized weights. Otherwise, it performs performance accounting based on the initial indicators and the corresponding initial weights. The target progress early warning module is used to determine whether to push early warning information to the preset responsible person based on the results of performance accounting and the indicator deviation coefficient used to reflect the degree of deviation between each indicator and the corresponding reference indicator. The performance accounting and improvement module is used to perform performance accounting update evaluation based on the parameters of the performance accounting process, and to determine whether there is a need to generate a performance accounting update plan, so as to improve the accuracy of performance accounting. The optimization of the aforementioned metrics is as follows: All initial indicators are sorted in descending order according to their priority reflecting their importance to the annual goals; If an initial indicator has a higher priority than the reference priority and its corresponding initial weight is less than the set core minimum weight, the preset personnel will be prompted to manually modify the initial weight; otherwise, each weight will be normalized. The initial weights corresponding to each initial indicator are split according to the set contribution ratio of each college; If the sum of the weights of the initial indicator across all colleges after splitting is not 1, then the weights of the initial indicator in each college are normalized to obtain optimized weights; otherwise, the split weights are directly set as optimized weights.

2. The university performance evaluation management system based on multi-source data fusion according to claim 1, characterized in that, The specific method for evaluating the rationality of multi-level indicator decomposition is as follows: The multi-path dependency conflict coefficient is obtained by comparing the number of repeated initial metrics with the total number of initial metrics. The multi-path dependency conflict coefficient is used to evaluate the complexity of the dependency relationship of the initial metrics in the multi-level decomposition process. The relative deviation between the total weight of the current level and the total weight of the next level is processed to obtain the level weight deviation coefficient, which is used to measure the degree of deviation between the current level and the next level in weight allocation. The relative deviation of the initial indicator summary results at the current level and the initial indicator summary results at the next level is processed to obtain the indicator summary result deviation coefficient. The initial indicator summary results represent the result obtained by weighting all initial indicators and their corresponding initial weights at the same level. The indicator summary result deviation coefficient is used to evaluate the difference between the current level and the next level in the indicator summary results. By introducing a reasonable evaluation and balancing weight, the multi-path dependency conflict index, the hierarchical weight deviation coefficient, and the index summary result deviation coefficient are subjected to inverse proportional assignment coupling processing to obtain the index decomposition reasonable index. The index decomposition reasonable index is used to quantitatively evaluate the rationality of the initial multi-level index decomposition and weight allocation process.

3. The university performance evaluation management system based on multi-source data fusion according to claim 2, characterized in that, The specific process for determining whether there is a need for indicator decomposition and optimization based on the obtained rationality assessment results is as follows: If the reasonable index of indicator decomposition is greater than the reasonable threshold of indicator decomposition, then no indicator decomposition optimization will be performed; otherwise, indicator decomposition optimization will be performed. Specifically, the indicator decomposition optimization includes: The initial metrics at each level are modeled as directed graphs, and depth-first search is used to detect whether circular dependencies exist. If a circular dependency exists, saving is prohibited, and the designated personnel are prompted to modify the dependency relationship; otherwise, initial weight optimization is performed. The initial weight optimization specifically involves: if the sum of the weights at the current level is not 1, then normalizing each weight; otherwise, prompting a designated person to manually modify the initial weights.

4. The university performance evaluation management system based on multi-source data fusion according to claim 3, characterized in that, The optimization of the indicator decomposition also includes: The initial weights of each initial indicator are sorted in descending order to obtain the minimum initial weight. The indicator with the minimum initial weight is set as the indicator to be judged, and the remaining initial indicators are set as indicators to be assigned. If the minimum initial weight is less than the minimum weight threshold, the weight of the indicator to be judged will be set to the minimum weight threshold, and the remaining weight will be allocated to the indicator to be allocated according to the weight ratio of each initial indicator; otherwise, no weight allocation will be performed.

5. The university performance evaluation management system based on multi-source data fusion according to claim 1, characterized in that, The specific process for determining whether to push the warning information to the preset responsible person is as follows: The initial indicators and their corresponding initial weights are assigned and coupled to obtain performance scores to measure the overall performance level of each object within a specific evaluation period. If the performance score is lower than the first reference performance score, an early warning message will be sent directly to the designated person in charge. If the performance score is less than the second reference performance score, and the number of times it exceeds the second reference performance score is greater than the preset number, then a warning message is directly pushed to the preset responsible person. Otherwise, it is determined whether to push a warning message to the preset responsible person based on the indicator deviation coefficient. The first reference performance score is less than the second reference performance score. The indicator deviation coefficient represents the result obtained by performing relative deviation processing on the initial indicator and the corresponding reference indicator. The specific process for determining whether to push early warning information to preset responsible persons based on the indicator deviation coefficient is as follows: If there is an initial indicator whose deviation coefficient is greater than the first indicator deviation threshold, then the warning information will be directly pushed to the preset responsible person. If there is an initial indicator whose deviation coefficient is greater than the second indicator deviation threshold, and the number of times it exceeds the optimized indicator deviation threshold is greater than the preset number, then a warning message will be sent directly to the preset responsible person, and the first indicator deviation threshold is greater than the second indicator deviation threshold. The warning information will not be sent to the designated responsible persons within the preset cooling-off period after the warning information is sent.

6. The university performance evaluation management system based on multi-source data fusion according to claim 1, characterized in that, The specific method for updating and evaluating performance based on parameters from the performance accounting process is as follows: The number of initial indicators whose coverage rate, which reflects the degree of implementation of the initial indicators at the teacher level, is greater than the coverage rate of the reference indicators is compared with the total number of initial indicators to obtain the annual target coverage consistency coefficient. The indicator coverage rate is the result obtained by comparing the number of teachers to whom each initial indicator is allocated with the total number of teachers. The performance score deviation coefficient is obtained by performing relative deviation processing on each performance score and the corresponding reference performance score obtained from the teacher level, and the average performance score deviation coefficient is obtained by averaging all the performance score deviation coefficients. Each initial indicator is paired with the Pearson correlation coefficient to obtain the initial indicator correlation coefficient. Indicators whose initial indicator correlation coefficient is greater than the reference correlation coefficient are set as redundant indicators. The number of redundant indicators obtained by statistics is compared with the total number of initial indicators to obtain the indicator redundancy. By introducing updated evaluation balance weights, the annual target coverage consistency coefficient, the average deviation coefficient of performance score, and the indicator redundancy are assigned values. Then, the assigned annual target coverage consistency coefficient is inversely proportionally calculated and coupled with the assigned average deviation coefficient of performance score and indicator redundancy to obtain the performance accounting update index. The performance accounting update index is used to quantify the probability of updating the performance accounting scheme.

7. The university performance evaluation management system based on multi-source data fusion according to claim 6, characterized in that, The determination is made as to whether there is a need to generate a performance accounting update plan; If the performance accounting update indicator is greater than the first performance accounting update threshold, a performance accounting update plan will be generated and sent to the preset performance management personnel. If the performance accounting update indicator is greater than the second performance accounting update threshold and the performance accounting update cycle is greater than the preset performance accounting update cycle, a performance accounting update plan will be generated and sent to the preset performance management personnel; otherwise, performance accounting will continue to be carried out according to the current performance accounting update plan. Determine whether a rejection notice for the performance accounting update plan has been received from the preset performance management personnel. If so, continue to perform performance accounting according to the current performance accounting version. Otherwise, update the performance accounting version according to the generated performance accounting update plan, map historical indicators to match the new performance accounting version, retain historical versions, and clear the cache of historical indicators whose retention time is longer than the preset retention time. The aforementioned performance accounting update scheme is as follows: If the average target achievement rate is less than the minimum target achievement rate threshold, the preset performance management personnel will be directly prompted to update the performance accounting scheme. If the average target achievement rate is greater than the maximum target achievement rate threshold, the performance accounting update indicators and performance input weight correction factors are mapped together to obtain the weight correction factor, and the initial weights are adjusted according to the weight correction factor. Otherwise, no performance accounting update plan is generated, and feedback is sent to the preset performance management personnel.

8. The university performance evaluation management system based on multi-source data fusion according to claim 7, characterized in that, The aforementioned performance accounting update scheme is as follows: When the standard deviation of the target achievement rate is greater than the standard deviation of the reference target achievement rate, or the target correlation coefficient is less than the reference target correlation coefficient, the initial indicators and annual targets are input into the Pearson correlation analysis to obtain the target correlation coefficient, and the current weights are adjusted according to the output target correlation coefficients. The calculation time window is obtained by mapping the performance accounting update indicators and the performance score input time window mapping set.

9. A university performance evaluation management method based on multi-source data fusion, applied to the university performance evaluation management system based on multi-source data fusion as described in any one of claims 1-8, characterized in that, include: S1. Obtain initial indicators reflecting the completion of teaching tasks, teachers' research activities, and teaching quality feedback of designated universities, and allocate the initial indicators to the corresponding implementers according to the target management hierarchy, which includes the university level, college level, and teacher level. S2, assign weights to the initial indicators of each target management level to obtain corresponding initial weights to reflect the contribution of each initial indicator to the annual target, conduct a multi-level indicator decomposition rationality assessment to measure the rationality of the initial indicator decomposition and weight allocation process, and determine whether there is a need for indicator decomposition optimization based on the obtained rationality assessment results. If so, perform performance accounting based on the optimized indicators used to update performance accounting and the corresponding optimized weights; otherwise, perform performance accounting based on the initial indicators and the corresponding initial weights. S3, based on the results of performance accounting and the indicator deviation coefficient used to reflect the degree of deviation between each indicator and the corresponding reference indicator, determine whether to push early warning information to the preset responsible person; S4. Based on the parameters of the performance accounting process, conduct a performance accounting update assessment and determine whether there is a need to generate a performance accounting update plan in order to improve the accuracy of performance accounting.

Images