Data processing method and system suitable for organization construction business evaluation system

Abstract

The invention discloses a data processing method and system suitable for an organization construction business evaluation system. The method comprises the following steps: acquiring organization construction data corresponding to an organization construction business to be evaluated; sequentially processing the organization construction data to obtain to-be-processed data; processing the to-be-processed data based on the constructed multi-dimensional evaluation model to obtain evaluation score information, the multi-dimensional evaluation model being designed to dynamically adjust the weight of each target dimension corresponding to the to-be-evaluated organization construction service according to an adaptive weight distribution mechanism, and calculating the evaluation score information according to the adjusted weight; generating a corresponding visual result log; and respectively sending the visualization result logs to the corresponding terminals. According to the data processing method and system suitable for the organization construction service evaluation system, the digitization and intelligentization process of organization construction service evaluation is promoted by designing a perfect flow of the data processing method, and then more comprehensive, objective and efficient comprehensive evaluation is achieved.

Description

Technical Field

[0001] The present invention relates to the field of computer processing technology, and in particular to a data processing method and system suitable for an organization construction business evaluation system. Background Art

[0002] In the existing technology, organizational construction business evaluation not only requires offline manual operations, such as filling out paper forms and manually summarizing data, which is inefficient and prone to errors, but also relies on the subjective experience of the evaluators, resulting in inaccurate evaluation results and making it difficult to achieve intelligent data management and monitoring of the organizational construction business evaluation system.

[0003] It can be seen that how to design a data processing method suitable for the organization construction business evaluation system to promote the digitalization and intelligentization of the organization construction business evaluation has become a technical problem that needs to be urgently solved by technical personnel in this field. Summary of the Invention

[0004] The present invention provides a data processing method and system suitable for an organization construction business evaluation system, which abandons the drawbacks of the existing technology that relies on offline and subjective experience. By designing a complete process of the data processing method suitable for the organization construction business evaluation system, the digitalization and intelligentization of the organization construction business evaluation are promoted, thereby achieving a more comprehensive, objective and efficient comprehensive evaluation.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a data processing method applicable to an organization construction business evaluation system, comprising: Obtain the organizational construction data corresponding to the organizational construction business to be evaluated; The organizational construction data is sequentially cleaned, formatted, standardized, and filled to obtain data to be processed; Processing the data to be processed based on the constructed multi-dimensional evaluation model to obtain evaluation score information corresponding to the organizational development business to be evaluated, wherein the multi-dimensional evaluation model is designed to dynamically adjust the weights of various target dimensions corresponding to the organizational development business to be evaluated according to an adaptive weight allocation mechanism, and calculating the evaluation score information using the adjusted weights; Perform intelligent analysis on the assessment score information and generate a corresponding visual result log; The visualization result log is sent to at least one of the following terminals: a platform terminal, an examiner's mobile terminal, or a data storage cloud.

[0006] As one preferred solution, obtaining the organizational construction data corresponding to the organizational construction business to be evaluated includes: Use data crawler technology to obtain the first organizational construction sub-data corresponding to the organizational construction business to be evaluated; Import the second organizational construction sub-data corresponding to the organizational construction business to be evaluated through a pre-set automated interface; Perform content recognition and analysis on the data in the local database to obtain the third-party organization construction sub-data corresponding to the organization construction business to be evaluated; The first organization construction sub-data, the second organization construction sub-data, and the third organization construction sub-data are integrated and analyzed to obtain the organization construction data.

[0007] As one preferred solution, the organization construction data is sequentially cleaned, formatted, standardized, and supplemented to obtain the data to be processed, including: Obtaining first data to be processed after the data standardization processing; If it is detected that the missing value of the first data to be processed does not reach the preset threshold, the first data to be processed is filled in based on the mean gap filling algorithm; If it is detected that the first data to be processed meets a preset skewed distribution condition, the first data to be processed is filled in based on a median filling algorithm; If it is detected that the data volume of the first data to be processed is within a preset range, the first data to be processed is supplemented based on a prediction algorithm.

[0008] As one of the preferred solutions, the construction of the multi-dimensional evaluation model includes: Determining organizational development business objective data and scoring dimensions corresponding to the organizational development business objective data; Based on the hierarchical analysis method, the organizational construction business target data is quantified to obtain each target level; Converting the weights of each target level into the weights of the corresponding scoring dimensions according to a weighting matrix; Dynamically adjusting each of the weights based on an adaptive weight allocation mechanism; The multi-dimensional evaluation model is constructed using the dynamically adjusted weights.

[0009] As one preferred solution, the adaptive weight allocation mechanism is based on which each weight is dynamically adjusted, including: adjusting each of the weights based on a linear regression model; Feedback correction is performed on the effects of the adjusted weights according to the Q-learning model.

[0010] As one preferred solution, the adaptive weight allocation mechanism is based on which each weight is dynamically adjusted, including: Repeatedly modifying each of the weights in time series based on the Holt-Winters exponential smoothing method; Perform difference analysis on the corrected errors until the difference analysis results meet the preset requirements.

[0011] As one preferred solution, the organization construction data includes historical organization construction data and real-time organization construction data; After obtaining the assessment score information corresponding to the organizational construction business to be assessed, the method further includes: Performing big data analysis and processing on the real-time organization construction data; Predicting trend forecast information corresponding to the organizational construction business to be assessed based on the big data analysis and processing results, the historical organizational construction data, and the assessment score information; Based on the trend forecast information, the weights in the multi-dimensional evaluation model are strengthened and revised.

[0012] As one preferred solution, the formats of the visualization result log include at least a document, a bar chart, a pie chart, and a radar chart.

[0013] As one of the preferred solutions, the intelligent analysis of the assessment score information to generate a corresponding visual result log includes: Performing accuracy and integrity checks on the assessment score information in sequence to obtain data to be processed; Creating various chart data corresponding to the data to be processed; Based on the large language model technology, generating natural language text data to be corrected that matches each of the chart data; Performing correction processing on the natural language text data to be corrected to obtain natural language text data; Generate a corresponding visualization result log using the natural language text data and the chart data.

[0014] Another embodiment of the present invention provides a data processing system suitable for an organization construction business evaluation system, comprising: The acquisition module is used to obtain the organizational construction data corresponding to the organizational construction business to be evaluated; A processing module is used to sequentially perform data cleaning, data formatting, data standardization, and data filling processing on the organizational construction data to obtain data to be processed; an evaluation module, configured to process the data to be processed based on a constructed multi-dimensional evaluation model to obtain evaluation score information corresponding to the organizational development business to be evaluated, wherein the multi-dimensional evaluation model is designed to dynamically adjust the weights of the various target dimensions corresponding to the organizational development business to be evaluated according to an adaptive weight allocation mechanism, and calculate the evaluation score information using the adjusted weights; An analysis module, configured to perform intelligent analysis on the assessment score information and generate a corresponding visual result log; The sending module is used to send the visualization result log to at least one of the following terminals: a platform terminal, an examiner's mobile terminal, or a data storage cloud.

[0015] Compared with the prior art, the embodiments of the present invention have the following advantages: Abandoning the drawbacks of existing technologies that rely on offline and subjective experience, a complete data processing method suitable for the organization construction business evaluation system is designed. Specifically, the organization construction data corresponding to the organization construction business to be evaluated is first obtained to provide accurate data support for subsequent data processing; then, in order to ensure data quality, targeted data cleaning, data formatting, data standardization and data filling processing are carried out in sequence; then a multi-dimensional evaluation model is constructed. The multi-dimensional evaluation model in the present invention is no longer a simple fixed weight calculation, but dynamically adjusts the weights according to an adaptive weight allocation mechanism, thereby effectively improving the generalization ability of the multi-dimensional evaluation model and the accuracy of the output results; then the relevant result log is generated and finally sent to different data terminals. The whole process integrates data collection, processing and analysis, and no longer relies on the subjective experience of the evaluator, realizing a comprehensive digital process, and providing scientific decision-making support for organization construction and human resource management. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flowchart of a data processing method applicable to an organization construction business evaluation system in one embodiment of the present invention; Figure 2 This is a logic block diagram of a data processing method applicable to an organization construction business evaluation system in one embodiment of the present invention; Figure 3 This is a structural block diagram of a data processing system applicable to an organization construction business evaluation system in one embodiment of the present invention; Reference numerals: Among them, 11, acquisition module; 12, processing module; 13, evaluation module; 14, analysis module; 15, sending module. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0018] In the description of this application, the terms "first," "second," "third," etc. are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first," "second," "third," etc. may explicitly or implicitly include one or more of the features. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0019] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be a communication between the two components. The terms "vertical", "horizontal", "left", "right", "up", "down" and similar expressions used herein are for illustrative purposes only, and do not indicate or imply that the device or component referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0020] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this application have the same meanings as those commonly understood by those skilled in the art. The terms used in this specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. Those skilled in the art will understand the specific meanings of the above terms in this application in specific circumstances.

[0021] It should be noted in advance that, although the existing general assessment management system is applicable to a variety of business scenarios, the evaluation method mainly relies on the subjective evaluation of the assessors, which makes it difficult to fully reflect the complex actual situation. In addition, there is a general problem that the assessment dimension is single, the specific needs of the organization construction assessment are not optimized enough, there is a lack of in-depth understanding of the organization construction business, it cannot effectively support multi-level and multi-department collaborative assessment, and it cannot flexibly meet the needs of different organizations for comprehensive assessment of the organization construction business. In view of this, the present invention designs a data processing method suitable for the organization construction business assessment system to solve the problems of insufficient data integration, subjective evaluation deviation, poor dynamic adaptability and high computational complexity in the existing technology, so as to achieve a more comprehensive, objective and efficient comprehensive evaluation.

[0022] Specifically, an embodiment of the present invention provides a data processing method applicable to an organization construction business evaluation system, see Figure 1 , Figure 1 The flowchart of the data processing method applicable to the organization construction business evaluation system in one embodiment of the present invention is shown, including steps S1 to S5, which are specifically as follows: S1. Obtain the organizational construction data corresponding to the organizational construction business to be evaluated; S2. Perform data cleaning, data formatting, data standardization, and data filling processing on the organizational construction data in order to obtain data to be processed; S3. Processing the data to be processed based on the constructed multi-dimensional evaluation model to obtain evaluation score information corresponding to the organizational development business to be evaluated, wherein the multi-dimensional evaluation model is designed to dynamically adjust the weights of the various target dimensions corresponding to the organizational development business to be evaluated according to an adaptive weight allocation mechanism, and calculate the evaluation score information using the adjusted weights; S4. Perform intelligent analysis on the assessment score information and generate a corresponding visual result log; S5. Send the visualization result log to at least one of the following terminals: a platform terminal, an examiner's mobile terminal, or a data storage cloud.

[0023] The embodiment of the present invention mainly realizes efficient, statistical and fair organizational construction evaluation by automatically integrating multi-dimensional data, dynamically adjusting the evaluation model and reducing the computational complexity. For ease of understanding, the following description is made in conjunction with the organizational construction business evaluation system.

[0024] Optionally, the organization construction business evaluation system may include the following modules: (1) Data processing module: Integrates the middle-end data of the organization construction domain and the human resources domain, such as personnel certificate information, personnel titles, responsibility information, etc., and cleans and pre-processes the data.

[0025] (2) Evaluation model module: Based on the integrated data and the actual situation of the organization, a dynamically adjusted evaluation model is constructed to provide multi-dimensional evaluation results.

[0026] (3) Dynamic prediction module: Based on historical data and real-time data, it uses machine learning algorithms to predict the future trend of organizational construction work and dynamically adjust the evaluation criteria.

[0027] (4) User feedback module: By combining the assessment scores and dynamic performance of personnel and using natural language generation (NLG) technology, the system can automatically generate personalized feedback reports for each person, helping them identify strengths, discover problems, and improve themselves.

[0028] For details, see Figure 2 , Figure 2 The following is a logical block diagram of a data processing method for an organizational development business evaluation system, according to one embodiment of the present invention. The first stage involves data acquisition (collection), which primarily utilizes automated interfaces and data import functions to acquire various organizational development-related data. Data sources include, but are not limited to, organizational development work data: work reports, activity records, and task completion status.

[0029] In the above embodiment, the data acquisition method can be through data crawler technology, automatic interface or local database import. It only needs to be integrated and analyzed after obtaining the sub-data of each organization construction, which will not be described in detail here.

[0030] After obtaining the above data, a series of processing steps are required to ensure data quality and consistency, including but not limited to data cleaning, data formatting, data standardization, and data imputation. Data cleaning involves cleaning the collected data, removing invalid data, filling missing values, correcting format inconsistencies, deleting duplicates, completing missing data, and addressing outliers. Data formatting converts data from different data sources to a unified format, such as date or number format. Data standardization normalizes data according to required standardized indicators to ensure effective comparison of data across different dimensions. Data imputation involves imputing missing key data using mean, median, or predictive algorithms. If the data is evenly distributed and has few missing values, using mean imputation is a simple and effective option. If the data contains extreme values ​​or a skewed distribution, using median imputation is more appropriate. When the data quality is high and the data volume is large, using predictive algorithms for imputation can achieve better results.

[0031] The multi-dimensional evaluation model in the embodiment of the present invention is used to score and evaluate the work of organizational construction. The main contents include: Dimension setting: The system will set corresponding evaluation indicators according to the various dimensions of organizational construction work. Each dimension may contain multiple detailed indicators. Weighted evaluation: The weights of different dimensions and detailed indicators can be configured according to the actual needs of the organization. The system dynamically adjusts the weight of each dimension and finally calculates the comprehensive assessment score of the personnel. Finally, different dimensions are scored by comprehensively using mathematical models and algorithms to ensure the detectability and scientific nature of the scoring process. This will be explained in detail later.

[0032] Furthermore, after the multi-dimensional assessment model is executed, embodiments of the present invention also perform trend forecasting. By analyzing historical and real-time data, machine learning algorithms are used to predict future trends in organizational development work. Evaluation criteria are dynamically adjusted based on the forecast results, ensuring more accurate, timely, and forward-looking assessments. Furthermore, if the forecast indicates that a particular task may present significant problems, the system automatically increases the weight of the relevant dimension to strengthen the evaluation of that dimension. In other words, based on trend forecast information, the weights in the multi-dimensional assessment model are strengthened and revised, allowing the multi-dimensional assessment model to be adjusted according to the organization's current work status, ensuring the model's timeliness.

[0033] Based on the multi-dimensional evaluation model and the dynamic scoring after the adjustment of the system, a comprehensive score is calculated for each organization, and a detailed evaluation report is generated. The report content shall at least include: Comprehensive score: showing the comprehensive score of the organizational construction work, as well as the scores of each dimension. Strengths and weaknesses analysis: based on the scoring results, the system analyzes the outstanding performance and areas that need improvement in the organizational construction work, and puts forward suggestions for improvement. Visualization results: using data visualization technology such as bar charts, pie charts, radar charts, etc., to quickly display the evaluation results to help users understand. The above results will be presented through the user interface or report generation system and sent to the relevant terminals.

[0034] In this embodiment of the present invention, dimension scores include periodic scoring data for each dimension. Historical scores and trends include trend data for individual scores for each dimension over a historical period, facilitating analysis of changing patterns in dimension performance. External factors, such as organizational level changes, may influence weight adjustments. Target-to-actual deviations analyze the discrepancy between an individual's historical scores and their expected targets to identify potential scoring deviations, which serve as a basis for weight optimization.

[0035] The goal of dynamic weight optimization is to enable the model to flexibly adjust the weights of various dimensions through continuous learning and adjustment based on changes in individual performance, feedback, and environmental factors, thereby achieving a precise, personalized, and fair scoring system. The following section details the dynamic weight optimization method, covering the application of time series analysis, machine learning models, reinforcement learning, and other optimization techniques.

[0036] Definition and modeling of organizational goals (i.e., organizational construction business goal data): Organizational goals typically include annual goals, long-term mission goals, and short-term mission goals. In order to connect with the evaluation system, these goals need to be quantified and converted into specific measurable indicators. These goals can be divided into several levels: Annual goals: such as staff participation in activities, etc.

[0037] Long-term mission objectives: such as strengthening the ideological construction of personnel and improving leadership capabilities.

[0038] Task objectives: tasks for a specific period, such as improving the effectiveness of organizational construction work.

[0039] Then, the organization's goals need to be accurately modeled and quantified into evaluation dimensions. Technical implementation methods include: Target hierarchical modeling (Analytical Hierarchy Process AHP): Goal Hierarchy: Based on the organization's overall mission objectives, break down the goals into different sub-goals and construct a goal tree. For example, annual goals can be broken down into "personnel ideological development" and "work performance," etc.

[0040] AHP application: Quantify the priorities of goals through expert review or historical data. By calculating the weight of each goal level, the system can identify which goals are more important to the long-term development of the organization.

[0041] Target quantitative indicators: For each sub-goal, clearly defined scoring dimensions are used.

[0042] Use a goal weight matrix: Use a weighting matrix to convert the weights of each goal level into weights for the scoring dimensions. In this process, AHP provides a clear quantitative framework to ensure that the scores for each dimension are consistent with the organizational goals.

[0043] The adaptive weight allocation mechanism in the embodiment of the present invention includes the following contents: (1) Generation of training dataset: Matching goals to scoring dimensions: First, determine how each assessment dimension maps to the organization's goals.

[0044] Label generation: Generate labels based on the priority of the organizational goal. The label of each sample will represent the relative importance of the organizational goal (e.g., 0.8 indicates a high priority for the goal, while 0.2 indicates a low priority).

[0045] (2) Use regression model to optimize weights: Linear regression: Using a linear regression model, we input the historical scores and target weights for each dimension, output a composite score, and ultimately optimize the dimension weights using a loss function (such as mean squared error (MSE)). By training on historical data, the model learns how to adjust the weights based on the performance of the scoring dimension.

[0046] Gradient Boosted Decision Tree (GBDT): Using tree models like GBDT, we can automatically handle complex relationships between features and evaluate the contribution of different dimensions to the score. Each tree "splits" its decision based on historical data across different dimensions, ultimately outputting the weights for each dimension.

[0047] (3) Weight optimization and adaptive adjustment: L1 / L2 regularization: To prevent weights from being overly dependent on a particular feature, L1 or L2 regularization methods can be used to ensure weight smoothness. L1 regularization can produce a sparse solution, forcing the weights of unimportant features to be zero; L2 regularization balances the weights of various features to avoid overfitting.

[0048] Adaptive objective function: During each optimization process, the objective function structure is adjusted based on changes in organizational goals. Specifically, the weights in the objective function can be dynamically adjusted based on the priority of organizational goals, thereby aligning the weight adjustment process with changes in goals.

[0049] (4) Reinforcement learning and real-time weight updates Reinforcement learning can be used to automate the process of optimizing weight distribution, ensuring that the system can adaptively adjust weights based on real-time feedback.

[0050] Optionally, feedback correction is performed on the effects of the adjusted weights according to a Q-learning model.

[0051] The state space of the Q-learning model: The scores of each evaluation dimension serve as the system state (for example, the performance scores of individuals), forming the state space. The degree to which the performance of each dimension matches the organizational goals determines the current state.

[0052] Action space of the Q-learning model: Weight adjustment is the "action" of reinforcement learning, that is, the system chooses whether to increase or decrease the weight of the dimension based on the current state (that is, the match between the performance of the scoring dimension and the organizational goals).

[0053] Reward function of the Q-learning model: The reward is calculated based on the similarity between the performance of the dimension and the organizational goal.

[0054] The Q-learning algorithm includes the following: State: The dimension scores and weights at the end of each scoring cycle. Action: Adjust the weights of each dimension. Reward: The degree to which the adjusted score matches the actual performance, or the accuracy of the score. Q-value: Indicates the value of taking an action in the current state. The update formula is: in, is the learning rate, is the discount factor, It’s an instant reward. is the maximum Q value of all actions in the next state.

[0055] Q-learning involves calculating the change in weight for each dimension based on new evaluation data, using a historical Q-value table (a table of weights that align dimensions with objectives). If the adjusted weight of a dimension better reflects the organization's objectives, the system increases that weight. Through exploration and utilization, the system continuously adjusts its strategy between existing historical data and new data to achieve the optimal weight distribution.

[0056] Optionally, after each evaluation cycle, the system incrementally trains the reinforcement learning model using the new evaluation data, updating the Q-value table to adjust dimension weights in real time. This incremental learning allows the system to dynamically update the weight allocation strategy based on changes in organizational goals over time, avoiding the impact of retraining on real-time performance.

[0057] Optionally, based on the Holt-Winters exponential smoothing method, dynamic weight adjustment based on time series analysis can be performed as follows: Time series analysis can identify trends, seasonal fluctuations, and unusual behavior in scoring data over time, providing a basis for dynamic weight adjustments. By predicting the performance of each dimension over a period of time, the system can adjust weights based on trends.

[0058] The Holt-Winters exponential smoothing method is a commonly used time series forecasting method, suitable for handling data with trends and seasonal fluctuations. In an organizational development scoring system, dimensions such as work performance may change over time. This method can capture these changing trends and adjust the score weights accordingly.

[0059] The Holt-Winters model includes two types: additive and multiplicative. The additive model is suitable for time series data with relatively stable fluctuations and is suitable for processing relatively stable score data such as personnel ratings.

[0060] The model consists of three main components: Level: Indicates the basic horizontal position of time series data.

[0061] Trend: Indicates the rate of change of data (i.e., the rate of increase or decrease).

[0062] Seasonality: A pattern that represents periodic fluctuations in data.

[0063] The system currently uses an additive model, and its time series update formula is: level( ): trend( ): Seasonality ( ): predict( ) in, is the observed value (i.e. the rating value of the current period), It's horizontal. It's a trend. It's seasonal. 、 、 is the smoothing parameter and m is the seasonal period.

[0064] The specific applications are as follows: In organizational development assessments, historical score trends are crucial for dynamic weight adjustments. The system can apply the Holt-Winters model to each assessment dimension (such as work performance) for predictions and dynamically adjust weights based on the predicted results. The following are the specific steps for this application: Data preparation: First, the system collects the score data of each evaluation dimension in the past period of time (for example, the past 12 months). Assume that the score data of the work performance dimension is , where t is the time point.

[0065] Model training: Input historical score data into the Holt-Winters model. The model will output three key parameters: (smoothing coefficient), (trend smoothing coefficient), (Seasonal smoothing coefficients). These parameters will be used to calculate the trend value for each dimension.

[0066] Smoothing process: The model calculates the long-term trend and seasonal fluctuations of the current dimension score through weighted averaging.

[0067] Trend prediction: Using the obtained trend parameters, the system can predict the future trend of the dimension scores. For example, if the scores of work performance are on an upward trend, the system will predict that the scores will continue to rise in the future.

[0068] Predicted Scores and Trend Adjustment: Based on the predicted score trends, the system will automatically determine the weight of the current dimension. If the performance score continues to rise, the system can increase the weight of this dimension to make its impact on the scoring results more prominent.

[0069] Weight Increase: If a dimension's score trend is predicted to rise, the system will increase its weight to emphasize its importance in the assessment. For example, the weight of "Work Performance" will be increased from the initial value of 10% to 15% to reflect its importance at the current stage.

[0070] After each scoring process, the system will obtain the actual scoring results. By comparing the predicted scores with the actual scores, the system can calculate the error and make feedback adjustments based on the error, thereby improving the model's prediction accuracy and real-time adaptability.

[0071] Error calculation: The system calculates the error by comparing the difference between the predicted score and the actual score. Specifically, the error is: Error = Actual score - Predicted score For example, if the predicted score for the job performance dimension is 80 and the actual score is 75, then the error is -5.

[0072] The system uses the error to fine-tune the model. This process is mainly achieved in two ways: Parameter update: Based on the error, the three smoothing parameters (α, β, and γ) of the Holt-Winters model are fine-tuned. By adjusting these parameters, the model can better adapt to the actual data, thereby improving the accuracy of the prediction.

[0073] Dynamic weight adjustment: If the error is large, it indicates that the current prediction may be inaccurate. The system will fine-tune the dimension weight based on the error value. For example, when the error is large, the system may temporarily reduce the weight of the dimension to avoid over-reliance on the score of a certain dimension.

[0074] After each scoring, the system will adjust the weights based on the error feedback, so that the prediction model can continuously optimize itself and improve the accuracy of prediction and scoring. The system feedback process includes the following steps: Evaluate the error size: If the error is small, the system considers the current weight distribution to be reasonable and continues to maintain the existing dimension weights.

[0075] If the error is large, the system will re-evaluate whether the current weight is reasonable, especially for dimensions where trend forecasts deviate, the system will dynamically adjust the weight.

[0076] Weight fine-tuning: Positive fine-tuning: When the prediction deviation is small and the trend dimension performs well, the system can slightly increase the weight of the dimension to further highlight its role in the scoring.

[0077] Negative fine-tuning: When the prediction error of a certain dimension is large, the system will reduce the weight of this dimension to reduce its impact on the scoring results until the system converges to a suitable weight distribution through the feedback mechanism.

[0078] Real-time adjustments: Because organizational development assessments involve comprehensive scoring across multiple dimensions, a real-time feedback mechanism ensures the system can dynamically adjust the weighting of each dimension based on current circumstances. This allows the weighting of dimensions like work performance to be continuously optimized based on historical scoring trends and real-time feedback data.

[0079] After each scoring, the system fine-tunes the parameters by comparing the difference (error) between the predicted score and the actual score to ensure the real-time adaptability of the model.

[0080] In an embodiment of the present invention, by combining the assessment scores and dynamic performance of personnel, the system can provide each person with detailed and specific feedback to help them identify strengths, discover problems, and improve themselves. The system uses natural language generation (NLG) or large language model technology to enable the organizational construction comprehensive evaluation system to automatically generate personalized feedback reports. Combined with chart data, it not only improves the efficiency of feedback, but also enhances the readability and incentive effect of the feedback content.

[0081] In another embodiment of the present invention, the system first needs to obtain structured data from various dimensions of organizational development assessment (such as learning, work, activities, etc.), including: Dimension score: The score value of each dimension (such as 0-100 points, or classification scores of "excellent", "qualified", "unqualified"). Dimension performance: Data such as course grades, activity participation, and task completion rate. Historical performance data: Past assessment records of personnel, providing comparison and trend analysis. Personalized characteristics: Such as the nature of the personnel's work and the personnel's development stage (new employee, veteran employee), etc. After data cleaning and standardization, this data is formed into a unified format for the NLG system to call.

[0082] The logic for generating personalized feedback mainly includes: 1. Template selection based on scores: The system selects an appropriate feedback template based on the person's scores or evaluations in various dimensions (e.g., excellent academic performance, less participation in activities).

[0083] 2. Dynamic Template Construction: For each dimension, the system can generate targeted feedback based on the specific score. For example, if the learning dimension score is high, the feedback content can be more motivational and praise-oriented; while if the activity dimension score is low, improvement suggestions can be generated.

[0084] For example, based on the number of statistical participation, the system calculates the participation ratio of each person in each dimension and dynamically adjusts the weight value in the evaluation model according to the ratio. The specific steps are as follows: Ratio calculation: For each person, calculate their participation ratio in each dimension. For example: Person A's course learning ratio is: 5 times / 10 times = 50%.

[0085] Person A's theme day ratio is: 3 times / 10 times = 30%.

[0086] The volunteer service ratio of person A is: 2 times / 10 times = 20%.

[0087] Weighting: The system dynamically adjusts the weights of each dimension in the evaluation model based on participation rates. For example, if a participant's participation rate in course learning is high, the system may increase the weight of that dimension to reflect their outstanding performance in that area. If a participant's participation rate in volunteer service is low, the system may decrease the weight of that dimension to avoid significantly impacting their overall score.

[0088] In this example, to ensure the feedback's tone is appropriate (both fair and motivational), the NLG system uses a sentiment analysis module to evaluate the feedback's tone. Sentiment analysis adjusts the feedback's tone based on the individual's score, ensuring it's both constructive and encouraging. For example, based on the individual's overall score and scores on each dimension, the system generates a personalized feedback log, helping them identify strengths, identify weaknesses, and improve themselves. The specific steps are as follows: Feedback content generation: Strength Recognition: The system generates praise feedback based on the individual's high scores in various dimensions. For example, "You have performed outstandingly in the course, having participated five times. We recommend that you continue to maintain this performance and deepen your study." Problem Discovery: The system generates improvement suggestions based on individuals' low scores in various dimensions. For example, "Your participation in volunteer services is low. We recommend increasing your participation in group activities to enhance organizational cohesion." Feedback strategy suggestions: Organizational Development Rotation Strategy: Based on an employee's performance across various dimensions, the system recommends different types of organizational development activities to achieve comprehensive development. For example, "You've been heavily involved in theme days. We recommend increasing your participation in volunteer activities to enhance your overall capabilities." Organizational construction business configuration strategy: Based on the performance of personnel, the system recommends that they invest more energy in certain dimensions.

[0089] The core of NLG is to convert structured data (scores, performance, etc.) into natural language text. There are two main ways to build an NLG model: 1. Template-based NLG: Fixed template + filler data: Combines a preset template with actual data to generate feedback content. The template usually includes a general feedback framework containing preset phrases or sentence patterns. For example: "Your academic performance is [score], and your performance is [excellent / good / acceptable / needs improvement]**. We recommend that you continue to maintain and further study [course content]." "Your participation in activities is low, and the current score is **[activity score]**. We recommend increasing participation in group activities to improve organizational cohesion." 2. Deep Learning-Based NLG (e.g., GPT-like models): Data-driven text generation: Leveraging deep learning models (e.g., pre-trained generative models like the GPT series and T5) to learn from large amounts of employee feedback data, this model generates more natural and personalized feedback. The model generates natural language text based on employee performance data and historical feedback.

[0090] In the organizational development assessment system, feedback isn't limited to text; it can also be multimodal. As needed, in addition to textual feedback, the system can also incorporate charts (such as bar charts and radar charts) to display employee scores across various dimensions.

[0091] The generated feedback report (log) will be output in a customizable format to ensure that the needs of different people or organizations are met. Feedback can be output in the following ways: Online platform terminal display: Personalized feedback is displayed through the personnel evaluation management system, and personnel can view it at any time.

[0092] Email or SMS notification to personnel's mobile terminal: The system sends email or SMS based on personnel's preferences to provide feedback content.

[0093] It should be noted that in cases where paper reports are required, the system supports generating personalized reports in PDF format for organizational use. In addition, relevant logs need to be uploaded to the data storage cloud for storage.

[0094] As can be seen from the above, the embodiment of the present invention mainly performs dynamic correction on the weight, which will be explained below in comparison with the static weight of a single dimension.

[0095] Limitations of single-dimensional static weights: ① Fixed weights: In a single-dimensional static weight model, the weights for each dimension are fixed and cannot be adjusted based on actual performance. For example, the weight for course learning is fixed at 0.5, and this weight remains unchanged regardless of how a person's performance in that dimension changes.

[0096] ② Lack of real-time performance: Static weighting models cannot respond to changes in personnel performance in real time. For example, if a person's course performance improves significantly during a certain period, the static weighting model cannot adjust the weight value in time, resulting in delayed evaluation results.

[0097] ③ Lack of foresight: Static weighting models cannot predict future performance trends and can only evaluate based on historical data. For example, if a person's volunteer performance shows an upward trend, the static weighting model cannot adjust the weight in advance, resulting in a lack of foresight in the evaluation results.

[0098] Advantages of dynamically adjusting weights: ① Dynamically adjust weights: The dynamic prediction module can dynamically adjust the weights of each dimension based on historical and real-time data. For example, if a person's performance in a course is on the rise, the system can appropriately increase the weight of that dimension to reflect their outstanding performance in that area.

[0099] ②Real-time response to changes: The dynamic prediction module can respond to changes in personnel performance in real time. For example, if a person's volunteer service performance improves significantly during a certain period, the system can immediately adjust the weight of that dimension to ensure the real-time evaluation results.

[0100] ③ Forward-looking prediction: The dynamic prediction module can predict future performance trends and adjust weight values ​​based on the prediction results.

[0101] Example: Dynamic weight adjustment of course learning dimensions.

[0102] Background: Person A's participation in courses has gradually increased over the past six months, and his performance has shown an upward trend.

[0103] Static weight model: The weight value of course learning is fixed at 0.5. No matter how the performance of person A changes, the weight value remains unchanged.

[0104] Dynamic Prediction Module: The system uses time series analysis to predict that student A's performance in course learning will continue to improve. The system dynamically adjusts the weight of course learning from 0.5 to 0.6.

[0105] Results: Personnel A's scores on course learning improved significantly, and the evaluation results more accurately reflected his actual performance.

[0106] Another embodiment of the present invention provides a data processing system suitable for an organization construction business evaluation system. Figure 3 , Figure 3 The following is a structural block diagram of a data processing system applicable to an organization construction business evaluation system in one embodiment of the present invention, which includes: An acquisition module 11 is used to acquire the organizational construction data corresponding to the organizational construction business to be evaluated; The processing module 12 is used to sequentially perform data cleaning, data formatting, data standardization, and data filling processing on the organizational construction data to obtain data to be processed; An evaluation module 13 is configured to process the data to be processed based on a constructed multi-dimensional evaluation model to obtain evaluation score information corresponding to the organizational development business to be evaluated, wherein the multi-dimensional evaluation model is designed to dynamically adjust the weights of the various target dimensions corresponding to the organizational development business to be evaluated according to an adaptive weight allocation mechanism, and calculate the evaluation score information using the adjusted weights; The analysis module 14 is used to perform intelligent analysis on the assessment score information and generate a corresponding visual result log; The sending module 15 is used to send the visualization result log to at least one of the following terminals: a platform terminal, an examiner's mobile terminal, or a data storage cloud.

[0107] The data processing method and system for an organization construction business evaluation system according to an embodiment of the present invention have the following beneficial effects: (1) By building a multi-source data integration mechanism, we have achieved deep integration of data from multiple fields, such as organizational development and human resources. By introducing an improved fuzzy comprehensive evaluation algorithm and combining it with machine learning technology, we have optimized the subjective evaluation data. Specific technical means include fuzzy matrix optimization: clustering analysis of subjective evaluation data through machine learning algorithms to identify and eliminate abnormal evaluations and reduce the impact of personal bias; dynamic weight adjustment: combining historical data and real-time data to dynamically adjust the weight of each factor to ensure the objectivity and scientific nature of the evaluation results.

[0108] (2) It effectively solves the problems of insufficient data integration, subjective evaluation bias, poor dynamic adaptability and high computational complexity in the existing technology, abandons the drawbacks of the existing technology that relies on offline and subjective experience, and designs a complete data processing method process suitable for the organization construction business evaluation system. Specifically, first obtain the organization construction data corresponding to the organization construction business to be evaluated to provide accurate data support for subsequent data processing; then, in order to ensure the quality of the data, carry out targeted data cleaning, data formatting, data standardization and data filling processing in turn; then construct a multi-dimensional evaluation model. The multi-dimensional evaluation model in the present invention is no longer a simple fixed weight calculation, but dynamically adjusts the weight according to the adaptive weight allocation mechanism, thereby effectively improving the generalization ability of the multi-dimensional evaluation model and the accuracy of the output results; then generate relevant result logs and finally send them to different data terminals. The whole process integrates data collection, processing and analysis, no longer relies on the subjective experience of the evaluator, realizes a comprehensive digital process, and provides scientific decision support for organization construction and human resource management.

[0109] The above-described embodiments merely illustrate several implementations of the present invention, and while their descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A data processing method suitable for an organization construction business evaluation system, characterized in that: include: Obtain the organizational construction data corresponding to the organizational construction business to be evaluated; The organization construction data is sequentially cleaned, formatted, standardized and supplemented to obtain data to be processed; The data to be processed is processed based on the constructed multi-dimensional evaluation model to obtain the evaluation score information corresponding to the organization construction business to be evaluated, wherein the multi-dimensional evaluation model is designed to dynamically adjust the weights of each target dimension corresponding to the organization construction business to be evaluated according to an adaptive weight allocation mechanism, and calculate the evaluation score information with the adjusted weights; Perform intelligent analysis on the assessment score information and generate a corresponding visual result log; The visualization result log is sent to at least one of the following terminals: a platform terminal, an examiner's mobile terminal, or a data storage cloud.

2. The data processing method applicable to the organization construction business evaluation system according to claim 1, characterized in that: The obtaining of the organizational construction data corresponding to the organizational construction business to be evaluated includes: Obtain the first organizational construction sub-data corresponding to the organizational construction business to be evaluated through data crawler technology; Import the second organization construction sub-data corresponding to the organization construction business to be evaluated through a preset automation interface; Perform content recognition and analysis on the data in the local database to obtain the third-party organization construction sub-data corresponding to the organization construction business to be evaluated; The first organization construction sub-data, the second organization construction sub-data and the third organization construction sub-data are integrated and analyzed to obtain the organization construction data.

3. The data processing method applicable to the organization construction business evaluation system according to claim 1, characterized in that: The data cleaning, data formatting, data standardization and data supplementation processing are performed on the organization construction data in sequence to obtain the data to be processed, including: Acquire the first data to be processed after the data standardization processing; If it is detected that the missing value of the first data to be processed does not reach the preset threshold, the first data to be processed is filled in based on the mean gap filling algorithm; If it is detected that the first data to be processed meets a preset skewed distribution condition, the first data to be processed is filled in based on a median filling algorithm; If it is detected that the data volume of the first data to be processed is within a preset range, the first data to be processed is supplemented based on a prediction algorithm.

4. The data processing method applicable to the organization construction business evaluation system according to claim 1, characterized in that: The construction of the multi-dimensional evaluation model includes: Determine organizational construction business target data and scoring dimensions corresponding to the organizational construction business target data; Based on the hierarchical analysis method, the business target data of the organization construction is quantified to obtain each target level; Converting the weights of each target level into the weights of the corresponding scoring dimensions according to a weighting matrix; Based on an adaptive weight allocation mechanism, dynamically adjust each of the weights; The multi-dimensional evaluation model is constructed with the dynamically adjusted weights.

5. The data processing method applicable to the organization construction business evaluation system as claimed in claim 4, characterized in that: The method of dynamically adjusting each weight based on the adaptive weight allocation mechanism includes: Adjusting each of the weights based on a linear regression model; Feedback correction is performed on the effects of the adjusted weights according to the Q-learning model.

6. The data processing method applicable to the organization construction business evaluation system according to claim 4, characterized in that: The method of dynamically adjusting each weight based on the adaptive weight allocation mechanism includes: Repeatedly modifying each of the weights in time series based on the Holt-Winters exponential smoothing method; Perform difference analysis on the corrected errors until the difference analysis results meet the preset requirements.

7. The data processing method applicable to the organization construction business evaluation system according to claim 1, characterized in that: The organization construction data includes historical organization construction data and real-time organization construction data; After obtaining the assessment score information corresponding to the organization construction business to be assessed, the method further includes: Performing big data analysis and processing on the real-time organization construction data; Predicting trend prediction information corresponding to the organization construction business to be evaluated based on the big data analysis and processing results, the historical organization construction data and the assessment score information; Based on the trend prediction information, the weights in the multi-dimensional evaluation model are enhanced and revised.

8. The data processing method applicable to the organization construction business evaluation system according to claim 1, characterized in that: The format of the visualization result log includes at least a document, a bar chart, a pie chart and a radar chart.

9. The data processing method applicable to the organization construction business evaluation system as claimed in claim 8, characterized in that: The intelligent analysis of the assessment score information to generate a corresponding visual result log includes: The accuracy and integrity of the assessment score information are sequentially checked to obtain data to be processed; Creating various chart data corresponding to the data to be processed; Based on the large language model technology, generating natural language text data to be corrected that matches each of the chart data; Performing correction processing on the natural language text data to be corrected to obtain natural language text data; Generate a corresponding visualization result log using the natural language text data and the chart data.

10. A data processing system suitable for an organization construction business evaluation system, characterized in that: include: An acquisition module is used to acquire the organizational construction data corresponding to the organizational construction business to be evaluated; A processing module, used to sequentially perform data cleaning, data formatting, data standardization and data filling processing on the organization construction data to obtain data to be processed; An evaluation module is used to process the data to be processed based on the constructed multi-dimensional evaluation model to obtain the evaluation score information corresponding to the organization construction business to be evaluated, wherein the multi-dimensional evaluation model is designed to dynamically adjust the weights of each target dimension corresponding to the organization construction business to be evaluated according to an adaptive weight allocation mechanism, and calculate the evaluation score information with the adjusted weights; An analysis module, used to perform intelligent analysis on the assessment score information and generate a corresponding visual result log; The sending module is used to send the visualization result log to at least one of the following terminals: a platform terminal, an assessor's mobile terminal or a data storage cloud.

Images