Mine workshop performance assessment management method

Abstract

The invention relates to the technical field of performance appraisal management, in particular to a mine workshop performance appraisal management method, which comprises data acquisition and preprocessing, dynamic performance evaluation, task difficulty adjustment, an incentive mechanism and performance feedback and optimization. Optimizing factor weights and predicting employee performance based on a multi-factor weighting model in combination with a neural network and reinforcement learning, and calculating performance scores in real time; the task difficulty is quantified by applying an analytic hierarchy process and fuzzy comprehensive evaluation, and real-time adaptive optimization is analyzed in combination with employee ability; a market supply and demand and collaborative filtering algorithm is adopted to adjust an integral value and a customization scheme, and an award standard is analyzed and optimized according to cost effectiveness and performance sensitivity; a report is generated by means of a visual library, suggestions are generated by means of text mining, and through prediction and causal analysis early warning, a management strategy adjustment scheme is recommended according to reinforcement learning, and feedback optimization is performed; according to the invention, the performance of the mine workshop can be comprehensively, accurately and dynamically assessed and managed, the operation efficiency and the enthusiasm of employees are improved, and sustainable development of enterprises is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of performance appraisal management, and in particular to a mine workshop performance appraisal management method. Background Art

[0002] Performance appraisal management of mining workshops is crucial to enterprise operations, but traditional methods have many drawbacks. In data collection and processing, the data source is single and mostly relies on simple equipment parameter records. There is a lack of collection of multi-dimensional equipment status and employee work details. In addition, data processing is insufficient, there are noise and missing values, and there is no effective integration and analysis, which cannot provide an accurate data basis for appraisal. In terms of performance evaluation, it focuses on a single production indicator, ignores key factors such as task diversity, complexity, and employee teamwork and innovation ability, and the evaluation is one-sided, affecting employee enthusiasm and enterprise efficiency. When adjusting the difficulty of tasks, the task allocation lacks scientificity, does not consider the adaptation of tasks and employee skills, and there is no real-time monitoring and dynamic adjustment mechanism, which affects employee performance and production efficiency. The incentive mechanism is single and lacks personalization, is not closely related to performance appraisal, and lacks long-term incentives. It cannot meet the diverse needs of employees and is not conducive to the long-term development of the enterprise. In the performance feedback and optimization link, the feedback lacks analysis and suggestions, and the management strategy adjustment has no data-driven basis. It is mostly empirical response, and it cannot achieve continuous optimization of the system. In short, traditional methods cannot meet the modern management needs of mining enterprises, and innovative performance appraisal management methods are urgently needed.

[0003] Therefore, a mine workshop performance appraisal management method is proposed to address the above problems. Summary of the invention

[0004] The purpose of the present invention is to provide a mine workshop performance assessment management method to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A mine workshop performance appraisal management method comprises the following steps:

[0007] Step S1, data collection and preprocessing: collect work data through sensor fusion, text analysis and employee self-assessment weights, and store it after multiple filling and robust standardization;

[0008] Step S2, dynamic performance evaluation: Based on a multi-factor weighted model, neural networks and reinforcement learning are combined to optimize factor weights and predict employees' future performance, and the performance score output is calculated in real time;

[0009] Step S3, task difficulty adjustment: using analytic hierarchy process and fuzzy comprehensive evaluation to quantify task difficulty, combined with employee capability analysis, relying on linear programming and gradient descent algorithm to adapt and optimize task difficulty in real time;

[0010] Step S4, incentive mechanism: using market supply and demand and collaborative filtering algorithms to adjust the points value and customize the points plan, optimizing the reward standards based on cost-effectiveness and performance sensitivity analysis, and realizing points redemption and data feedback;

[0011] Step S5, performance feedback and optimization: Generate performance reports with the help of visualization library and use text mining to generate improvement suggestions, warn of performance changes through prediction and causal analysis, and use reinforcement learning to recommend management strategy adjustment plans and provide feedback for optimization.

[0012] As a preferred solution, in data collection and preprocessing, data collection includes:

[0013] Sensors are installed on key equipment in the mine workshop, and the Kalman filter algorithm is used to fuse sensor data to obtain basic data related to equipment operation and production. At the same time, natural language processing (NLP) technology is used to analyze employee work record texts, and the key information in the texts is quantified and added to the work data. Based on the linear regression algorithm, the weight of employee self-evaluation in the overall data collection is determined according to the consistency between the employee's past self-evaluation and actual performance, and the employee work data is obtained. The employee work data is then integrated in a unified format and transmitted to the data preprocessing server with the help of big data message queue technology; among them, a linear regression model is constructed ,in, is the actual performance data set, A collection of employee self-evaluation data. is the intercept, is the regression coefficient, is the error term, and the least squares method is used to estimate and To determine the weight of employee self-assessment;

[0014] Data preprocessing includes:

[0015] On the data preprocessing server, the multiple imputation method MICE is used to fill in the data with missing values. Multiple prediction models are established based on the data association relationship to ensure data integrity. The robust standardization RobustScaler method is used to standardize the data. The calculation formula is: ,in, is the original data, is the standardized data, Representation data the median of Representation data The interquartile range of the preprocessed data is stored in a distributed database.

[0016] As a preferred solution, in dynamic performance evaluation, the multi-factor weighted model construction and optimization process includes:

[0017] Determine the task type, difficulty, employee historical performance, teamwork effect, and innovation results as the key factors for performance evaluation. Let these factors be , based on expert experience and historical data analysis, the weights of each factor are initially set ,and represents the total number of key factors for performance evaluation, It is an index variable used to traverse all key factors and build a multi-factor weighted model; performance score The calculation formula is:

[0018] ,in, is the optimized weight, are the factor values ​​corresponding to the current working data;

[0019] Then, using the neural network algorithm, the fluctuations in employee performance and changes in the focus of the workshop business are taken as inputs. According to the conventional training and weight adjustment mechanism of the neural network, the weights of each factor in the weighted model are adjusted in real time through back propagation. At the same time, the XGBoost algorithm is used in combination with reinforcement learning to build a prediction model, which takes employee work tasks as "actions" and performance scores as "rewards". Based on the exploration learning mechanism of reinforcement learning and combined with XGBoost's ability to process historical performance data, the prediction model is optimized to accurately predict employees' future performance.

[0020] The implementation method of performance score calculation and output is as follows: input the pre-processed employee work data into the optimized multi-factor weighted model, combine it with the prediction model results, calculate the employee performance score in real time according to the established calculation logic, and output it to the subsequent modules through the API interface to provide data support for related operations.

[0021] As a preferred solution, the hierarchical analysis method in task difficulty adjustment includes:

[0022] Construct a hierarchical model and set the target layer as the task difficulty ,The criterion layer contains relevant factors, and the judgment matrix is ​​constructed by comparing the factors in the criterion layer pairwise. ,in, Indication factors Relative to The importance of Scaling method assignment;

[0023] Solving the Judgment Matrix by Eigenvector Method The largest characteristic root and its corresponding eigenvector , calculate the consistency index ,in, To determine the matrix order and check consistency, Less than average random consistency index When , the consistency is considered to be passed, and the normalized feature vector That is, the weight vector of each criterion layer factor ;

[0024] The fuzzy comprehensive evaluation method includes:

[0025] Set of comments , fuzzy evaluation is performed on the sub-factors under each criterion layer, and a fuzzy relationship matrix is ​​constructed. ,in, Indication factors For comments The membership degree, combined with the criterion layer factor weight vector , using the fuzzy comprehensive evaluation formula , and obtain the fuzzy comprehensive evaluation result vector of task difficulty ,in, represents the fuzzy synthesis operator, which is then converted into a specific task difficulty value through the defuzzification method. The defuzzification method is the centroid method, and the calculation formula is: ,in, for Middle elements, For comment collection Middle The quantitative value corresponding to each comment, The final task difficulty value is stored in the task management database as a reference for task difficulty adjustment; For comment collection The number of comments in the

[0026] Cluster analysis algorithms are used to classify employees into different ability levels based on their past performance and skill mastery. The similarities between employees are calculated and classified using the corresponding clustering algorithms. For each type of employee, the Bayesian network model is used to analyze the probability relationship between tasks of different difficulty levels and employees' positive work efficiency. The network structure and conditional probability are learned based on historical data to provide a reference for the reasonable allocation of tasks.

[0027] The specific strategy for real-time adjustment of task difficulty is as follows:

[0028] Based on the employee performance scores and task completion status obtained from the performance evaluation module, combined with the task difficulty evaluation model and the employee capability adaptation analysis results, the linear programming algorithm is used to calculate the task difficulty adjustment range suitable for the employee's current stage with a suitable objective function and satisfy the corresponding constraints, and the task allocation is adjusted accordingly;

[0029] When employees are performing tasks, the real-time emotional state of employees obtained through wearable devices and workshop video monitoring, as well as the work efficiency tracking curve drawn using time series analysis, are combined to judge the effect of task difficulty adjustment. If it does not meet expectations, the task difficulty is quickly adjusted based on the iterative update mechanism of the gradient descent algorithm to ensure that employees maintain a positive working state under tasks of appropriate difficulty, and relevant information is promptly fed back and updated.

[0030] As a preferred solution, in the incentive mechanism, the collaborative filtering algorithm includes:

[0031] Let the user set be , the item set is , the user's rating matrix for items is ( Indicates user For items Ratings of users, if the user has not rated, it is recorded as a missing value); Rating matrix based on users ,To calculate the similarity between users, the Pearson correlation coefficient method is commonly used, and the calculation formula is ,in, and are different users, Is a user and A collection of items that have all been rated. and The users are and Average rating of Indicates user and The similarity between Indicates user For items Ratings;

[0032] Find target users Most similar Users form a neighbor set , predict the target user by weighted average For unrated items Rating , use collaborative filtering algorithm to analyze employee-related personalized characteristics, refer to the past performance improvement of similar employees, customize personalized points acquisition rules and distribution plans for employees according to the above algorithm, and employees will obtain points according to their performance after completing the task, which will be stored in the points account database;

[0033] Adopt a simulation algorithm based on market supply and demand theory, and dynamically adjust the corresponding points value according to the basic principles of the algorithm according to the popularity of different rewards among employees and the popularity of redemption;

[0034] Dynamic adjustments to reward standards include:

[0035] From the perspective of enterprise cost-effectiveness, the cost-effectiveness analysis model is used in combination with the dynamic programming algorithm. Based on the enterprise's business objectives and financial status, the reward standards are optimized in real time according to the corresponding algorithm logic. The difference analysis and elasticity analysis statistical methods are used to analyze the sensitivity between performance scores and changes in reward standards, so as to fine-tune the reward standards corresponding to different performance intervals. Employees choose to redeem rewards on the reward redemption platform based on their points and standards. The incentive mechanism module feeds back the redemption records to other related modules for subsequent optimization analysis.

[0036] By comprehensively considering the costs and benefits of rewards, analyzing the relationship between performance and rewards, setting reward standards scientifically and reasonably, employees are encouraged to improve their performance, and at the same time, the incentive mechanism and the entire performance appraisal management system are coordinated and optimized.

[0037] As a preferred solution, in performance feedback and optimization, management strategy adjustment and optimization include:

[0038] Use time series prediction algorithms to predict trends in team and individual performance data, and combine causal analysis algorithms to explore potential causes of performance changes. When the probability of performance decline or potential problems is predicted, the early warning mechanism is triggered according to the set threshold and the early warning information is pushed to the workshop management personnel;

[0039] Based on the reinforcement learning algorithm, by simulating the performance changes after the implementation of different management strategies, the optimal management strategy adjustment plan is recommended to managers according to the reinforcement learning strategy recommendation mechanism. Managers make corresponding adjustments in the workshop management system accordingly. The adjusted performance change data is fed back to the system for further optimization of the parameters and models of each algorithm module, so as to realize continuous iterative optimization of the entire performance appraisal management system and improve the overall operational efficiency of the mine workshop.

[0040] It can be seen from the technical solution provided by the present invention that the mine workshop performance appraisal management method provided by the present invention has the following beneficial effects:

[0041] 1. Data collection and processing are more scientific and comprehensive:

[0042] Multi-source data fusion ensures accuracy: By installing multiple sensors on key equipment in the mine workshop, such as output, pressure, operating time, speed and other sensors, and using the Kalman filter algorithm to fuse data, more accurate basic data related to equipment operation and production can be obtained; for example, the fusion of output and pressure sensor data on mining equipment can accurately reflect the relationship between the actual working status of the equipment and production efficiency, avoid single sensor errors, and provide a reliable basis for subsequent assessments;

[0043] Text analysis to mine potential information: Natural language processing (NLP) technology is used to analyze employee work record texts, and task progress, quality feedback and other information are quantitatively supplemented into work data, making assessment data richer and more comprehensive. For example, quality problem descriptions mined from employee records can timely discover production risks and help with targeted improvements.

[0044] Employee self-assessment weights optimize data quality: Employee self-assessment weights are determined based on a linear regression algorithm and adjusted based on the consistency between past self-assessments and actual performance, which improves the effectiveness of self-assessment data in overall data collection; if an employee's self-assessment is highly consistent with actual performance, a higher weight is given so that the overall data can more truly reflect the employee's work situation;

[0045] 2. More accurate and dynamic performance evaluation:

[0046] Comprehensive consideration of multi-factor weighted model: A weighted model is constructed by integrating multiple factors such as task type, difficulty, employee historical performance, team collaboration effect, and innovation results to avoid the one-sidedness of single-factor evaluation; for example, when evaluating employee performance, both the amount of task completion (related to task type and difficulty) and the contribution in team collaboration (team collaboration effect) are considered, making the evaluation more comprehensive and objective;

[0047] Neural networks and reinforcement learning optimize weights to predict the future: Neural networks adjust factor weights in real time based on employee performance fluctuations and changes in workshop business priorities. Reinforcement learning optimizes the prediction model with work tasks as "actions" and performance scores as "rewards" to accurately predict employees' future performance. If the workshop has recently focused on product quality (business focus changes), the neural network will increase the weights of quality-related factors accordingly. Reinforcement learning simulates task scenarios to allow employees to optimize performance strategies in different task choices, achieving accurate evaluation and prediction.

[0048] 3. The task difficulty is more adaptable:

[0049] Quantitative science of analytic hierarchy process and fuzzy comprehensive evaluation: analytic hierarchy process builds a task difficulty assessment model to determine the weights of criterion layer factors such as skill level, time urgency, and resource dependence. Fuzzy comprehensive evaluation combines factor weights with fuzzy relationship matrix, and obtains accurate task difficulty values ​​after defuzzification, providing a scientific quantitative basis for task allocation. For example, for a complex, time-sensitive, and resource-demanding task, this method can accurately assess its difficulty and reasonably allocate it to employees with matching abilities.

[0050] Employee capability analysis enables personalized adaptation: Cluster analysis classifies employees based on their past performance and skill mastery, and the Bayesian network model analyzes the adaptation relationship between different types of employees and task difficulty, achieving personalized adaptation between tasks and employee capabilities, and improving employee work efficiency and overall production efficiency; for example, employees are divided into high, medium, and low capability levels, and tasks of appropriate difficulty are assigned to employees of different categories. High-ability employees take on high-difficulty tasks to unleash their potential, and low-ability employees gradually improve from tasks of appropriate difficulty, avoiding inefficiency caused by tasks that are too difficult or too easy;

[0051] 4. The incentive mechanism is more flexible and effective:

[0052] Dynamic adjustment of points value to meet demand: Based on the market supply and demand theory, the points value is dynamically adjusted according to the attention and redemption popularity of different rewards, so that the points system is more in line with employee needs and improves the incentive effect; for example, the points value of popular training opportunities increases as demand increases, guiding employees to work hard to obtain points while reasonably allocating corporate resources;

[0053] Collaborative filtering customizes personalized plans: Use collaborative filtering algorithms to analyze employees' personalized characteristics, and customize points acquisition rules and distribution plans based on the performance improvement of similar employees, so as to provide more targeted incentives. For example, for employees with a strong need for technical improvement, a plan is customized to allow them to obtain specific training reward points by completing technology-related tasks, so as to meet their personalized development needs and stimulate their work enthusiasm.

[0054] Optimize reward standards through cost-effectiveness and performance sensitivity analysis: From the perspective of cost-effectiveness, use relevant algorithms to optimize reward standards in real time, fine-tune rewards for different performance intervals through performance sensitivity analysis, set reward standards scientifically and reasonably, and achieve coordinated optimization of incentives and corporate benefits. For example, when analysis finds that performance improvement is sensitive to a certain reward, increase the intensity of the reward appropriately, motivate employees to improve performance to the greatest extent under the premise of controlling costs, and ensure the balance of input-output benefits of the enterprise at the same time.

[0055] 5. Performance feedback and optimization form a virtuous cycle:

[0056] Visual reports and improvement suggestions help employees improve: With the help of the visualization library, intuitive and interactive performance reports are generated, so that employees can clearly understand their own performance and how they compare with the team; using text mining technology to generate targeted improvement suggestions, help employees identify the direction of improvement and promote personal performance improvement; for example, employees find that their work efficiency is lower than the team average through radar chart comparison, and combine improvement suggestions to reasonably plan work processes, refer to excellent practices, and improve work efficiency;

[0057] Prediction and analysis warning to help management decision-making: Use time series prediction and causal analysis algorithms to predict performance changes and explore potential causes, and recommend management strategy adjustment plans based on reinforcement learning to help managers deal with problems in advance and optimize management decisions. For example, if a team's performance decline is predicted and analysis shows that it is due to increased task difficulty and insufficient training, managers can organize training and adjust task allocation based on the recommended plan to avoid worsening performance problems.

[0058] Feedback optimization loop continuously improves system performance: performance data after the implementation of the adjustment plan is fed back to the system, and each module optimizes its own algorithm and function accordingly, so as to achieve continuous iterative optimization of the performance appraisal management system and improve the overall operational efficiency of the mine workshop; for example, the performance evaluation module optimizes the scoring model according to the impact of rewards on performance, and the task difficulty adjustment module adjusts the task difficulty more accurately according to the improvement of employee capabilities, so that the entire system can continuously adapt to the actual operational needs of the workshop and continuously improve management level and production efficiency;

[0059] 6. Improve the overall management level and competitiveness of mine workshops:

[0060] Optimize resource allocation: Through accurate performance evaluation, reasonable task allocation and effective incentive mechanism, employees can be better matched with tasks, and resources such as equipment can be more reasonably used to avoid idle resources or excessive consumption, thus improving the efficiency of resource allocation in mine workshops; for example, highly skilled employees can be assigned to key and complex tasks to ensure efficient completion of tasks, while preventing low-skilled employees from wasting time and resources on difficult tasks;

[0061] Enhance employee enthusiasm and sense of belonging: A fair, scientific and personalized performance appraisal management system enables employees to receive accurate evaluation and reasonable rewards for their efforts. The incentive mechanism meets the different needs of employees, enhances their work enthusiasm and sense of belonging to the company, and reduces staff turnover. For example, when employees see that their performance improvement brings actual rewards and career development opportunities, they will be more actively engaged in work and have higher loyalty to the company.

[0062] Adapt to market changes and enterprise development: The system can timely adjust the assessment management strategy according to changes in the internal and external environment of the workshop (such as adjustments in business priorities, changes in market demand, etc.), continuously optimize operating processes and employee performance, and help mining companies maintain their competitiveness in market competition and achieve sustainable development. For example, when the market has higher requirements for product quality, the system can adjust the weights of performance evaluation factors to guide employees to focus on quality improvement, so that the company's products meet market demand and gain a competitive advantage.

[0063] The mine workshop performance appraisal management method of the present invention has comprehensively improved the performance appraisal management level of mine workshops through multi-faceted innovation and optimization, bringing significant advantages to mining enterprises in production operations, employee management, market competition, etc., and has high practical value and promotion significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 The present invention is a schematic diagram of the steps of a mine workshop performance appraisal management method. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0066] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0067] like Figure 1 As shown, an embodiment of the present invention provides a mine workshop performance assessment management method, comprising the following steps:

[0068] Step S1, data collection and preprocessing: collect work data through sensor fusion, text analysis and employee self-assessment weights, and store it after multiple filling and robust standardization;

[0069] Step S2, dynamic performance evaluation: Based on a multi-factor weighted model, neural networks and reinforcement learning are combined to optimize factor weights and predict employees' future performance, and the performance score output is calculated in real time;

[0070] Step S3, task difficulty adjustment: using analytic hierarchy process and fuzzy comprehensive evaluation to quantify task difficulty, combined with employee capability analysis, relying on linear programming and gradient descent algorithm to adapt and optimize task difficulty in real time;

[0071] Step S4, incentive mechanism: using market supply and demand and collaborative filtering algorithms to adjust the points value and customize the points plan, optimizing the reward standards based on cost-effectiveness and performance sensitivity analysis, and realizing points redemption and data feedback;

[0072] Step S5, performance feedback and optimization: Generate performance reports with the help of visualization library and use text mining to generate improvement suggestions, warn of performance changes through prediction and causal analysis, and use reinforcement learning to recommend management strategy adjustment plans and provide feedback for optimization.

[0073] In this embodiment, the specific operation steps of step S1 include:

[0074] Step S1-1, data collection:

[0075] Step S1-1-1, sensor data collection and fusion:

[0076] A variety of sensors are installed on key equipment in the mine workshop, such as production sensors and pressure sensors on mining equipment, and running time sensors and speed sensors on transportation equipment. Each sensor collects corresponding data in real time to form multiple sensor data sets, which are recorded as ;

[0077] The Kalman filter algorithm is used to fuse the data collected by each sensor to improve the accuracy and reliability of the data. Taking two sensors as an example (it can be expanded to multiple sensors in practice), assume that sensor 1 is at time The data collected is , sensor 2 at time The data collected is ,The fusion is completed through the following steps:

[0078] First determine the observation matrix (Set according to the sensor data characteristics and fusion requirements according to the Kalman filter algorithm requirements);

[0079] Calculate Kalman gain (Calculated according to the recursive formula of the Kalman filter standard, combined with the covariance and other parameters of the sensor data);

[0080] The final fused data Calculated as follows: ;

[0081] The data after merging all moments is combined into a set As basic data related to equipment operation and production;

[0082] Step S1-1-2, text data collection and quantification:

[0083] Employees input work-related text records through the workshop work record system. These records include task progress descriptions, quality feedback, problem records, etc.

[0084] Use natural language processing (NLP) technology for analysis. The specific operations are as follows:

[0085] Use a word vector model (such as the Word2Vec algorithm) to convert words in text records into vector representations. For each text record , generate the corresponding vector ;

[0086] The vector is quantized and converted according to the predefined keyword and value mapping rules. For example, if a keyword combination indicating high-quality task completion appears in the text, it is mapped to a value of 8 according to the rules; if a general completion status expression appears, it is mapped to a value of 5, etc., thereby converting the text record into quantized text data, which is used together with the previous sensor fusion data as the data content to be integrated;

[0087] Step S1-1-3: Determine the weight of employee self-assessment and synthesize the data:

[0088] Collect employees' past self-evaluation data and corresponding actual performance data, and set them as employee self-evaluation data sets and actual performance data collection ;

[0089] Building a linear regression model (in, is the intercept, is the regression coefficient, is the error term), and the least squares method is used to estimate and

[0090] The calculation formula of the least squares method is (here To include independent variables The design matrix, is the dependent variable vector), calculate and An estimated value of

[0091] according to The value of and the goodness of model fit (such as The weight of employee self-assessment is determined by indicators such as like Close to 1 mesh Higher, The value is relatively large, otherwise the value is small;

[0092] It is determined , the employee self-assessment data is combined with sensor fusion data and quantified text data according to weights to form a complete employee work data set covering task completion, work efficiency, work quality, equipment operation status, etc. ;

[0093] Step S1-1-4, data integration and transmission:

[0094] Integrate all types of collected data in a unified data format (such as JSON format) to ensure that the data structure is standardized and unified for subsequent processing;

[0095] With the help of big data message queue (such as Kafka) technology, the integrated data is transmitted to the data preprocessing server in real time to ensure the timeliness of the data so that it can enter the next step of preprocessing process in time;

[0096] Step S1-2, data preprocessing:

[0097] Step S1-2-1: Multiple imputation method (MICE) to fill missing values:

[0098] After the data arrives at the data preprocessing server, for the data matrix with missing values (This includes the employee work data set collected and integrated earlier , there may be some missing data) for processing;

[0099] For columns with missing values Every missing value in ( Indicates row index), multiple prediction models are established based on the correlation between data (for example, linear regression models and decision tree models are constructed using other non-missing columns as independent variables):

[0100] For each prediction model , use the training data (i.e. the non-missing part of the data) to train the model and obtain the missing values The predicted value of ;

[0101] go through After filling, missing values The final filling value Calculated by the following formula: ;

[0102] In this way, all missing values ​​in the data are filled to improve the completeness of the data;

[0103] Step S1-2-2, Robust Standardization (RobustScaler) processing data:

[0104] The Robust Scaler method is used to standardize the data after filling in missing values, so that the data conforms to the normal distribution characteristics and is insensitive to outliers.

[0105] For the original data , standardized data Calculated according to the following formula: ;

[0106] in, It’s data the median of It’s data The interquartile range (the difference between the upper quartile and the lower quartile) is calculated for each data element in the data set to obtain the standardized employee work data set. ;

[0107] Step S1-2-3, storing the preprocessed data:

[0108] The employee work data set after multiple imputation and robust standardization Stored in a distributed database (such as HBase) so that subsequent dynamic performance evaluation, task difficulty adjustment, incentive mechanism, performance feedback and optimization modules can successfully call on these data as the basic data source to carry out corresponding work;

[0109] Through the operation of the above-mentioned detailed step S1 and its sub-steps, the whole process from data collection to preprocessing is completed, providing a high-quality and standardized data foundation for the entire mine workshop performance appraisal management system, ensuring that subsequent links can perform corresponding analysis and decision-making based on accurate and reliable data.

[0110] In this embodiment, step S2: specific operation steps of dynamic performance evaluation include:

[0111] Step S2-1: Construction and initialization of multi-factor weighted model:

[0112] Step S2-1-1, determine the key factors of performance evaluation: Analyze the work characteristics and performance evaluation requirements of the mine workshop, determine the task type, difficulty, employee historical performance, teamwork effect, innovation results and other factors as the key factors of performance evaluation, and form a set of these factors, recorded as ;

[0113] Step S2-1-2: Preliminary setting of factor weights: Based on expert experience and analysis of historical performance data of the mine workshop, preliminary weights are set for each key factor to form a weight set. , while satisfying The conditions are used to build the initial multi-factor weighted model and performance score In the initial stage, the calculation is performed according to the following formula: , is an index variable used to traverse all key factors, starting from 1 to End, where It represents the total number of key factors for performance evaluation;

[0114] Step S2-2: Optimize factor weights using neural networks:

[0115] Step S2-2-1, data preparation and input determination: Collect the recent performance fluctuation data of employees (such as the trend and fluctuation range of performance scores in the past period of time) and the information on changes in the focus of workshop business (such as the production indicators and quality improvement directions that the workshop has focused on recently), and organize these data into a format suitable for neural network input to form an input vector ;

[0116] Step S2-2-2, neural network structure construction:

[0117] Build a neural network for deep learning, including input layer, hidden layer (determine the number of hidden layers and the number of nodes in each layer according to actual conditions, and set the number of hidden layer nodes to , the activation function selects a suitable function, such as ReLU function, etc., recorded as And the output layer:

[0118] For the calculation from the input layer to the hidden layer, the weight matrix from the input layer to the hidden layer is The bias vector is , the output of the hidden layer Calculated according to the following formula: ;

[0119] The calculation from the hidden layer to the output layer, assuming that the weight matrix from the hidden layer to the output layer is The bias vector is , the output of the output layer (i.e. the adjusted factor weight vector Obtained by the following formula: ;

[0120] Step S2-2-3, weight adjustment training process:

[0121] Define a suitable loss function (such as the mean square error loss function, denoted as ,in, is the actual factor weight label, is the weight value predicted by the neural network, is the sample size, is the sample index variable), the gradient is calculated according to the loss function through the back propagation algorithm, and the weight of the neural network is continuously adjusted based on optimization algorithms such as gradient descent , so that the predicted factor weights Continuously approach the true optimal weight, and adjust the weight of each factor in the weighted model in real time based on employee performance fluctuations and changes in workshop business priorities;

[0122] Step S2-3: Build a prediction model by combining reinforcement learning:

[0123] Step S2-3-1: Prepare training data and define environmental elements:

[0124] Collect a large amount of historical performance data from mine workshops and build a training data set. Set as ,in, is the employee feature vector (including employee basic information, past task completion data, skill mastery and other features), Score the corresponding performance and define the relevant elements in reinforcement learning;

[0125] State space: The various comprehensive states that employees are in before performing each task are defined as state space ,For example, status factors including current employee skill level, recent performance, and the difficulty of the tasks they face;

[0126] Action space: Treat each work task performed by an employee as an "action" in reinforcement learning to form an action space ,Different tasks can be regarded as different action choices;

[0127] Transition Probability: Determine the transition probability , indicating that in the state Take action Transfer to state The probability of the task being executed can be obtained based on the statistical analysis of historical data, reflecting the possibility of the state changing after the task is executed;

[0128] Reward function: Use performance score as "reward" to construct reward function , that is, employees are in the state Next action The performance score obtained after completing the corresponding task is used as reward feedback to measure the quality of behavior;

[0129] Step S2-3-2: Build the XGBoost model foundation:

[0130] The XGBoost algorithm is used to build a machine learning prediction model, and its objective function By the loss function (For example, the commonly used logarithmic loss function is used to measure the prediction performance score and real performance rating The difference between ) and the regularization term (used to prevent overfitting and ensure the generalization ability of the model), the specific formula is as follows: ,in, is the sample index;

[0131] Based on the training data set ,Train the XGBoost model to enable it to make preliminary predictions on employee performance;

[0132] Step S2-3-3, reinforcement learning optimization process:

[0133] Combined with the Markov decision process in reinforcement learning, the intelligent agent follows a certain strategy (Indicates that in the state Take action The probability of ) is explored and learned in past task scenarios (i.e., the state space and action space constructed by historical data):

[0134] Use common reinforcement learning algorithms such as Q-learning to update the value function (Indicates that in the state Take action The agent continuously tries different actions (performs different tasks) and learns the strategy of selecting the best action in different states based on the reward (performance score) feedback obtained. This helps to find out the best performance improvement strategy for employees in different situations, and further optimizes the prediction model built by XGBoost, so that it can more accurately predict the future performance of employees.

[0135] Step S2-4, performance score calculation and output:

[0136] Step S2-4-1, data input and score calculation:

[0137] The current work data of employees obtained through the previous data collection and preprocessing steps (covering task completion, work efficiency, work quality and other aspects) is input into the multi-factor weighted model optimized through the above steps. At the same time, combined with the results output by the prediction model optimized by XGBoost combined with reinforcement learning, according to the calculation logic of the multi-factor weighted model (i.e. ,at this time, is the optimized weight, are the factor values ​​corresponding to the current working data, represents the total number of key factors for performance evaluation, is an index variable used to traverse all key factors) and calculate the specific performance score of each employee in real time ;

[0138] Step S2-4-2: Data output transmission:

[0139] The calculated performance score And the scores of each factor (i.e. The relevant data (values) are output to the subsequent task difficulty adjustment module and incentive mechanism module in the form of API interface, providing data support for subsequent task difficulty adjustment, incentive measures formulation and other related operations, ensuring that the entire performance appraisal management process can be smoothly connected and continuously promoted;

[0140] Through the above detailed step S2 and its sub-step operations, it is realized to optimize the factor weights based on the multi-factor weighted model, combine neural network and reinforcement learning, and accurately predict the future performance of employees. Finally, the performance score is calculated and output in real time, providing a scientific and dynamic evaluation basis for the performance appraisal management of mine workshops.

[0141] In this embodiment, step S3: task difficulty adjustment specific operation steps include:

[0142] Step S3-1: Quantification and analysis of task difficulty:

[0143] Step S3-1-1: Use the analytic hierarchy process (AHP) to build a comprehensive assessment model for task difficulty:

[0144] Build a hierarchical model:

[0145] Clearly define the target level as task difficulty , represents the final task difficulty result to be evaluated;

[0146] Determine the factors included in the criterion layer, such as skill level (covering operational skill complexity, knowledge and skill requirements, etc.), time urgency , resource dependence As well as other relevant factors (such as safety risk level, etc.) included according to the actual situation of the mine workshop, a hierarchical structure is constructed to clearly show the hierarchical relationship between the factors;

[0147] Construct a judgment matrix:

[0148] Compare the factors in the criterion layer pairwise and construct a judgment matrix based on their relative importance ,in, Indication factors Relative to The importance of is assigned using a 1-9 scale; for example, if Compare Equally important, ;like Compare Slightly more important, ; And so on, until the construction of the entire judgment matrix is ​​completed;

[0149] Calculate the weight vector and check consistency:

[0150] Solving the Judgment Matrix by Eigenvector Method The largest characteristic root and its corresponding eigenvector Commonly used calculation methods such as the power method (which can be implemented with the help of mathematical calculation software);

[0151] Calculate consistency index ( is the judgment matrix order), and compared with the average random consistency index (You can obtain the corresponding standard form by querying the corresponding Under value value) for consistency check, when Less than When , the consistency of the judgment matrix is ​​considered to be passed. At this time, the eigenvector is normalized to obtain the weight vector of each criterion layer factor , the weight vector reflects the relative importance of each criterion layer factor in measuring task difficulty;

[0152] Step S3-1-2: Determine the task difficulty value using the fuzzy comprehensive evaluation method:

[0153] Determine the comment set:

[0154] Set up a comment set based on the common evaluation levels of task difficulty in mine workshops , Indicates the first Comment elements, such as , used to describe different levels of task difficulty;

[0155] Construct the fuzzy relationship matrix:

[0156] For each sub-factor under each criterion layer, fuzzy evaluation is performed to determine the impact of each sub-factor on the comment set. The membership degree of each comment in the fuzzy relationship matrix is ​​constructed ,in, Indication factors For comments The membership degree can be determined by expert scoring, historical data statistical analysis, etc. For example, a number of experts familiar with the mine workshop business are invited to evaluate the task difficulty level corresponding to each sub-factor, and the frequency of each comment is counted as the membership value;

[0157] Fuzzy comprehensive evaluation calculation:

[0158] Combined criterion layer factor weight vector Fuzzy Relation Matrix , using the fuzzy comprehensive evaluation formula Calculate ( Represents fuzzy synthesis operator, commonly used are main factor prominent type , weighted average wait, Indicates taking the big operation, represents the smaller operation), and obtains the fuzzy comprehensive evaluation result vector of task difficulty , represents the factor index in the factor set, Represents the index of a comment in the comment set;

[0159] Defuzzification:

[0160] The fuzzy comprehensive evaluation result vector is transformed into Converted into a specific task difficulty value, for example, using the center of gravity method, the calculation formula is (in, for Middle elements, For comment collection Middle The quantitative value corresponding to each comment, is the final task difficulty value; For comment collection The calculated task difficulty value is stored in the task management database to provide an accurate reference standard for subsequent task difficulty adjustment;

[0161] Step S3-1-3, employee capability analysis:

[0162] Employee classification (cluster analysis):

[0163] Collect employees’ past performance data (such as performance scores at each stage, task completion rate, etc.) and skill mastery data (including proficiency levels of various skills, skill certificate acquisition status, etc.) to form data sets. Let the performance data set be , the skill mastery data set is ( is the number of employees);

[0164] Calculate the similarity between employees. Common similarity calculation methods include Euclidean distance (in, Indicates Employee's eigenvalues, is a feature dimension, which includes data on relevant dimensions such as performance and skill mastery). Other appropriate similarity measurement methods can also be used according to actual conditions.

[0165] Use cluster analysis algorithms (such as K-Means algorithm) to divide employees into different ability levels according to the calculated similarity between employees, and update the cluster center through continuous iteration (in K-Means algorithm, the initial random selection The cluster centers are then reallocated according to the distance between the employees and the cluster centers, and the cluster centers are updated. The iterations are repeated until the cluster centers no longer change or the preset number of iterations is reached (such as the stopping condition), so that the similarity of employees in the same cluster is high and the similarity between different clusters is low. Finally, Employee competency categories;

[0166] Analyze the relationship between tasks and employee capabilities (Bayesian network model):

[0167] For each category of employees, the Bayesian network model is used to analyze the probability relationship between tasks of different difficulty levels and employees' positive work efficiency. The task difficulty variable is set as , employee effectiveness variables are ;

[0168] Based on the historical data of the mine workshop (including the work efficiency data generated after different difficulty tasks were assigned to different categories of employees in the past, such as work efficiency improvement, task completion quality, etc.), learn the structure of the Bayesian network (i.e., the dependency graph between variables) and the conditional probability table Through the reasoning mechanism of Bayesian network, the probability of positive work efficiency that tasks of different difficulty levels may bring to a certain category of employees is analyzed, providing a scientific reference for the subsequent reasonable allocation of tasks;

[0169] Step S3-2: Real-time adjustment strategy for task difficulty:

[0170] Step S3-2-1, determine the task difficulty adjustment range (linear programming):

[0171] Objective function setting:

[0172] According to the employee performance scores and task completion status obtained from the performance evaluation module, combined with the task difficulty evaluation model and the employee ability adaptation analysis results, set a suitable objective function, such as maximizing performance improvement or maximizing employee satisfaction; assuming that the employee performance score is , the current task difficulty is , the expected range of task difficulty for employee ability adaptation is (obtained through Bayesian network analysis, etc.), the objective function Can be set to (in, The performance score of the employee after the task difficulty is expected to be adjusted. is the estimated value of employee satisfaction, and is the weight coefficient set according to the importance attached by the enterprise);

[0173] Constraints determination:

[0174] Clarify the range constraints of task difficulty adjustment , ensuring the adjusted task difficulty Within the reasonable range of employee capability adaptation; at the same time, consider other practical constraints, such as the limitation of resources required for the task (for example, some difficult tasks require specific equipment resources. If the current workshop has limited resources, the allocation of such difficult tasks needs to be restricted), the matching of task time requirements and employees’ available time, etc., and add these constraints to the linear programming model in the form of mathematical expressions;

[0175] Solve the task difficulty adjustment value:

[0176] Solve the constructed linear programming model through a linear programming solver (such as common algorithms such as the simplex method) to calculate the optimal task difficulty adjustment value , which is the specific value of the task difficulty adjustment range that best suits the employee at the current stage. Task allocation is adjusted accordingly. For example, for employees whose performance scores are continuously at a high level and whose ability adaptation analysis shows that they are more suitable for high-difficulty tasks, the difficulty of their subsequent tasks can be appropriately increased. Conversely, for employees with poor performance, the difficulty of their tasks can be appropriately reduced.

[0177] Step S3-2-2, task difficulty adjustment effect monitoring and iterative optimization (gradient descent algorithm):

[0178] Monitoring data collection:

[0179] When employees are performing tasks, wearable devices (such as smart bracelets, which can obtain physiological data such as employee heart rate and blood pressure, indirectly reflecting the employee's emotional state and work pressure) and workshop video monitoring (using image recognition technology to analyze employees' facial expressions, body movements, etc. to determine their emotional state) are used to obtain the employee's real-time emotional state. Emotional recognition algorithms (such as classification models based on deep learning convolutional neural networks) are used to classify and judge the employee's emotional state into positive, neutral, negative, etc.

[0180] At the same time, the time series analysis method is used to draw the employee work efficiency tracking curve and collect the work efficiency related data of employees in the task execution process (such as output per unit time, product qualification rate, etc.). Suppose the time series data is ,in, Indicates at time The collected work efficiency data can be smoothed by moving average, exponential smoothing and other methods to construct a suitable time series model (such as autoregressive model : (in, is the autoregression coefficient, is the error term, Representing time series The past Hysteresis value, ) to analyze performance change trends;

[0181] Effect judgment and adjustment decision:

[0182] Based on the collected data such as employees' real-time emotional state and work efficiency tracking curve, determine whether the task difficulty adjustment has achieved the expected effect. For example, if the employee's emotional state continues to be negative and work efficiency is on a downward trend, it can be considered that the task difficulty adjustment has not achieved the expected effect and needs further optimization.

[0183] Gradient descent iterative adjustment:

[0184] If the task difficulty adjustment is judged to be less than the expected effect, the task difficulty is adjusted rapidly and iteratively based on the gradient descent algorithm. , define a suitable loss function (For example, the sum of squares of the difference between employee work efficiency and expected efficiency is used as the loss measurement indicator), according to the update formula of the gradient descent algorithm (in, is the current task difficulty, is the learning rate, which needs to be set reasonably according to the actual situation. is the loss function exist The task difficulty is iteratively updated to ensure that employees maintain a positive working state under tasks of appropriate difficulty, and the adjusted task allocation information is fed back to the employee's work terminal in a timely manner and the relevant task information in the task management database is updated, so as to continuously track and further optimize the task difficulty adjustment;

[0185] Through the above detailed step S3 and its sub-step operations, the quantification of task difficulty based on hierarchical analysis method and fuzzy comprehensive evaluation is realized, combined with employee ability analysis, and the task difficulty is adapted and optimized in real time by relying on linear programming and gradient descent algorithm, making the task allocation of mine workshops more scientific and reasonable, which helps to improve employee work performance and overall production efficiency.

[0186] In this embodiment, step S4: specific operation steps of the incentive mechanism include:

[0187] Step S4-1, points system design and management:

[0188] Step S4-1-1: Adjust the points value based on market supply and demand theory:

[0189] Data Collection and Analysis:

[0190] Collect data on the popularity of different rewards (such as cash rewards, training opportunities, promotion opportunities, physical prizes, etc.) among employees in the mine workshop and data on their redemption frequency. These data can be obtained through internal reward redemption platform records, employee questionnaire feedback, and other channels; for example, statistics on the number of times employees browsed each reward and the number of times they applied for redemption within a period of time, so as to reflect the degree of employees' demand for different rewards;

[0191] Simulating market supply and demand:

[0192] Based on the collected data, the point value corresponding to the reward is dynamically adjusted by simulating the principle of market supply and demand. For rewards with high demand (i.e., high attention and high redemption frequency), the corresponding point value is increased according to certain rules to make them relatively "more expensive". On the contrary, for rewards with low demand, their point value is reduced to make it easier for employees to redeem them by accumulating fewer points. For example, if a training opportunity has been favored by many employees recently and the number of redemption applications has increased, the corresponding point value can be increased at a predetermined rate (such as 10%) at regular intervals (such as weekly or monthly); and for certain physical prizes that have been ignored for a long time, their point value can be reduced accordingly by a certain rate.

[0193] Step S4-1-2: Customize the scoring scheme using collaborative filtering algorithm:

[0194] Construct the user-item rating matrix:

[0195] Let the employee set be Represents the user, and the reward set is Representing items, constructing a user's rating matrix for items ,in, Indicates user For items The score can be customized and quantified based on factors such as the employee's preference for rewards and satisfaction after past redemption. If an employee does not rate a reward, it will be recorded as a missing value. For example, an employee can give a high score (such as 5 points) to a reward that he or she is very much looking forward to, and a low score (such as 1 point) to a reward that he or she is not very interested in. Initially, some score data can be collected through questionnaires or employees' initial evaluations on the reward redemption platform to construct a matrix.

[0196] Calculate user similarity:

[0197] User-based rating matrix , use a suitable similarity calculation method to calculate the similarity between users, such as the Pearson correlation coefficient method, which is calculated as follows: ;

[0198] in, and They are two different users. Is a user and A collection of items that have all been rated. and The users are and Average rating of Indicates user and The similarity between Indicates user For items The similarity value between each pair of users is calculated by this formula to form a user similarity matrix, which is used to measure the similarity between employees in terms of reward preferences;

[0199] Find neighboring users and predict ratings:

[0200] For each target user , find the most similar user among all users users (i.e., neighboring users), forming a neighbor set ; Usually you can sort by similarity value from high to low, select the first Users as neighbors;

[0201] By using weighted average and other methods, the rating data of neighboring users is used to predict the target user. For unrated items Rating ,For example, the prediction formula based on the weighted average of user similarity is: ;

[0202] According to the predicted scores, we analyze the potential preference of employees for different rewards, and customize individualized points acquisition rules and allocation plans for each employee based on their work performance (such as performance scores, skill improvement, etc.), job requirements and other factors. For example, for employees with excellent performance and predicted high interest in a certain professional skill training reward, we develop a plan that can quickly accumulate the points required for the training reward by completing difficult tasks; for new employees, we design a way to obtain points and allocation ratios that focuses more on helping them get familiar with the work environment and obtain opportunities to improve basic skills;

[0203] Employee points acquisition and storage:

[0204] After completing various work tasks, employees will obtain corresponding points according to their performance scores, task difficulty completion coefficient and other factors according to the customized personalized points plan. For example, completing a difficult and high-quality task can earn more points, while routine tasks can earn relatively fewer points.

[0205] The points data earned by employees is stored in the employee points account database, which makes it convenient for employees to check their points balance and points acquisition details at any time, and also facilitates subsequent points redemption operations and data statistical analysis;

[0206] Step S4-2: Dynamic adjustment of reward standards:

[0207] Step S4-2-1, cost-benefit analysis:

[0208] Cost accounting:

[0209] For each reward, calculate the costs involved in detail, including direct costs (such as the amount of cash rewards, the cost of purchasing physical prizes, etc.) and indirect costs (such as the teacher fees involved in providing training opportunities, venue rental fees, and future salary increase costs corresponding to promotion opportunities, etc.), and summarize the various costs to form the total cost data of the reward;

[0210] Continuously track and record the cost changes of various rewards in different periods. For example, as market prices fluctuate, the cost of purchasing physical prizes may increase or decrease; as the level of training teachers improves, the cost of training may also increase accordingly, etc., and update cost data in a timely manner to ensure the accuracy of cost accounting;

[0211] Benefit evaluation:

[0212] Analyze the benefits brought to the company after the award is issued, mainly from the perspective of employee performance improvement, such as comparing the changes in performance scores, work efficiency improvements, and changes in loyalty to the company before and after the employee receives the award. By establishing a quantitative evaluation indicator system (such as setting indicators such as performance improvement rate and work efficiency growth rate) to specifically evaluate the benefits brought by the award;

[0213] At the same time, the positive impact of rewards on team collaboration atmosphere and employee innovation enthusiasm should be considered, and the overall benefits of rewards should be comprehensively evaluated. For example, whether the internal cooperation projects of the team are smoother and more efficient under the incentive of rewards, whether employees have put forward more innovative working methods or suggestions, etc., and these qualitative influencing factors should be converted into quantifiable data as much as possible and included in the scope of benefit evaluation;

[0214] Step S4-2-2, dynamic programming algorithm optimization reward standard:

[0215] Building a decision model:

[0216] Based on the data of cost-benefit analysis, combined with the business objectives of enterprises at different stages (such as increasing production in the short term, improving enterprise innovation capabilities in the long term, etc.) and financial conditions (such as budget constraints, cash flow conditions, etc.), a reward standard optimization decision model based on dynamic programming is constructed to define state variables (such as reward status corresponding to different performance intervals, enterprise financial status in different time periods, etc.), decision variables (such as the specific amount of rewards, the choice of reward type, etc.) and state transition equations (a mathematical relationship that describes how reward standards change over time and when performance changes);

[0217] Solving the optimal reward standard:

[0218] Solve the constructed model through dynamic programming algorithms (such as conventional solution methods such as the inverse method) to calculate the optimal reward standards under different circumstances such as different stages and performance ranges under the current business conditions and goals of the enterprise. For example, determine what combination of reward amounts and reward types within a certain performance range can maximize employee performance and overall benefits while meeting the enterprise's cost budget, and record and update these optimal reward standards as the actual implementation reward standards;

[0219] Step S4-2-3: Performance sensitivity analysis and fine-tuning:

[0220] Analyze the sensitivity of performance ratings and reward criteria to changes:

[0221] Use statistical analysis methods (such as difference analysis and elasticity analysis) to conduct in-depth research on the relationship between changes in performance scores and changes in reward standards, calculate the sensitivity coefficient of performance scores to changes in reward standards (for example, through regression analysis and other methods, establish a functional relationship between performance scores and reward standards, and derive the sensitivity coefficient), and understand the sensitivity of employee performance to reward adjustments in different performance ranges. For example, it is found that employees with performance in the medium and high level range are more sensitive to the increase in reward amount; while for employees with lower performance, they may pay more attention to the diversification of reward types rather than simply the increase in amount;

[0222] Fine-tune the reward criteria:

[0223] According to the results of sensitivity analysis, fine-tune the reward standards corresponding to different performance ranges. For performance ranges with high sensitivity, appropriately increase the reward intensity (such as increasing the reward amount, adding more attractive reward types, etc.); for ranges with low sensitivity, optimize the reward structure (such as adjusting the combination of reward types, setting hierarchical reward conditions, etc.), so that the reward standards are more scientific, reasonable and targeted, and can better motivate employees to improve their performance;

[0224] Step S4-3, points redemption and data feedback:

[0225] Step S4-3-1, points redemption operation:

[0226] Employees can choose to redeem corresponding reward items on the reward redemption platform established within the company based on the number of points they have accumulated and the current reward standards. The redemption process must follow the procedures set by the platform, such as logging into the personal account first, viewing the list of redeemable rewards, deducting the corresponding points after confirming the redemption, and generating a redemption record, etc., to ensure the standardization and traceability of the points redemption operation;

[0227] Step S4-3-2, data feedback mechanism:

[0228] The incentive mechanism module will promptly feed back the employee's points redemption records (including details such as redemption time, redeemed rewards, points consumed, etc.) and employee performance change data after the reward is issued (such as the fluctuation of performance scores within a period of time after the reward is redeemed, changes in work efficiency, etc.) to other related modules such as the performance evaluation module and the task difficulty adjustment module, so that each module can perform corresponding optimization analysis based on these feedback information. For example, the performance evaluation module can further optimize the performance scoring model based on the impact of rewards on performance; the task difficulty adjustment module can refer to the improvement of employee capabilities under reward incentives and more accurately adjust the task difficulty range, etc., to achieve coordinated optimization and continuous improvement of the entire performance appraisal management system;

[0229] Through the above detailed step S4 and its sub-steps, we can rationally design the points system through market supply and demand and collaborative filtering algorithms, scientifically optimize the reward standards based on cost-effectiveness and performance sensitivity analysis, and effectively complete the points exchange and data feedback, thereby building a complete and flexible incentive mechanism to help mine workshops better motivate employees and improve overall operational efficiency and performance levels.

[0230] In this embodiment, step S5: specific operations of performance feedback and optimization include:

[0231] Step S5-1: Performance report generation and improvement suggestions provision:

[0232] Step S5-1-1: Generate performance report with the help of visualization library:

[0233] Data collection and compilation:

[0234] Collect various aspects of employees' performance information from various relevant data sources (such as employee performance scores and task completion data output by the performance evaluation module, task allocation and difficulty data from the task difficulty adjustment module, and point acquisition and redemption records from the incentive mechanism module, etc.), and organize these data according to certain rules and formats to ensure data integrity and accuracy. For example, map different types of performance data of each employee to corresponding data fields to form a structured data set for subsequent visualization processing;

[0235] Select a visualization type:

[0236] According to the characteristics of the performance information to be displayed and the purpose of analysis, select the appropriate visualization chart type and use the D3.js visualization library to generate performance reports; for example:

[0237] If you want to show the relationship between various performance factors of employees (such as task completion quality, work efficiency, teamwork, etc.) and their proportion to the overall performance, you can choose a Sankey diagram, which uses lines of different widths and directions to intuitively reflect the relationship and weight between various factors;

[0238] If you want to compare the comprehensive performance of employees at different stages (such as monthly or quarterly) and the difference with the team average, it is more appropriate to use a radar chart. Map the performance indicator values ​​at different stages to the coordinate axis of the radar chart to clearly show the performance trend and relative position.

[0239] Visual report generation and presentation:

[0240] Use the relevant functions and syntax of the D3.js library to visualize the sorted data according to the requirements of the selected chart type to generate an intuitive and interactive performance report; for example, by writing JavaScript code, configure the chart's coordinate axis, data binding, style settings and other parameters to make the performance report present a good visual effect;

[0241] The generated performance reports are deployed to the management platform within the mine workshop. Employees and managers can log in to the corresponding accounts and conveniently view the performance reports of their respective interests on the web page. Employees can intuitively understand their own strengths and weaknesses in the work process and the comparison with the overall level of the team, providing a reference for self-improvement;

[0242] Step S5-1-2: Generate improvement suggestions using text mining:

[0243] Data preprocessing:

[0244] Collect performance data and employees' daily work records (such as task summaries, problem feedback, experience sharing and other text content) as data sources for text mining, clean these text data, remove noise information (such as redundant punctuation, spaces, repeated meaningless content, etc.), and perform basic natural language processing operations such as lexical analysis and part-of-speech tagging to convert the text into a format that is convenient for subsequent analysis, such as dividing the text into meaningful words and marking their parts of speech;

[0245] Topic extraction and key information mining:

[0246] Use text mining technology (such as topic model algorithms, such as LDA algorithm) to conduct in-depth analysis of pre-processed text data and automatically extract key topics and problem points. For example, use LDA algorithm to analyze employee work record texts and mine topic information such as "equipment failure affects task progress" and "communication and collaboration efficiency needs to be improved" that reflects common problems and key areas of concern in the work;

[0247] Combined with predefined business rules and key indicators, further filter and extract key information closely related to performance improvement. For example, for texts related to task completion, focus on key descriptions that affect performance, such as task overtime and substandard quality.

[0248] Improvement suggestion generation:

[0249] Based on the mined key topics and information, combined with the rules set by the expert system (these rules are based on the business experience and best practices of the mine workshop, such as stipulating that if "equipment failure affects task progress", it is recommended to strengthen daily equipment maintenance and emergency response plan formulation) and pre-trained natural language generation models (which can be trained based on a large number of existing performance improvement case texts to enable it to generate suggestions that conform to language logic and actual business scenarios), targeted and highly operational improvement suggestions are automatically generated for employees. For example, for the problem of low work efficiency in employee performance, specific suggestions such as "it is recommended to reasonably plan the work process, refer to the operating methods of excellent employees in the same position, prepare for tasks in advance, and avoid unnecessary waste of time" are generated, and these improvement suggestions are displayed in conjunction with the corresponding performance reports to assist employees in better improving their performance levels;

[0250] Step S5-2: Performance change warning and management strategy recommendation:

[0251] Step S5-2-1, performance change prediction:

[0252] Data preparation and prediction model selection:

[0253] Collect past performance data of teams and individuals (including performance scores in different time periods, detailed data of each performance factor, etc.), and organize these historical performance data into time series data format in chronological order. For example, for the performance score data of a single employee, form a similar time series, where Indicates at time performance rating value;

[0254] According to the characteristics of performance data (such as whether there are seasonal fluctuations, trend changes, etc.), select appropriate time series prediction algorithms, such as the ARIMA model (applicable to stationary time series, which makes predictions by performing operations such as differencing on the data to make it stationary), the LSTM model in deep learning (which has a good prediction effect for time series with complex nonlinear characteristics and long-term dependencies), etc., and build corresponding prediction models;

[0255] Model training and prediction:

[0256] Use historical performance data to train the selected forecasting model. For example, in the ARIMA model, by estimating the model's autoregressive order, difference order, and moving average order, the optimal parameter values ​​are determined using methods such as maximum likelihood estimation so that the model can fit the historical data. In the LSTM model, configure the appropriate network structure (such as the number of hidden layer nodes and layers), and train and optimize the model using back-propagation algorithms.

[0257] Apply the trained model to predict future performance data. Input the corresponding feature data at the current time point (such as recent performance, external environmental factors, etc., which can be appropriately selected and preprocessed according to model requirements) and output the predicted value of performance changes in a certain period of time in the future (such as the next month or quarter). For example, predict that a team's performance score may show a downward trend in the next quarter.

[0258] Step S5-2-2: Cause and effect analysis to find out the reasons for performance changes:

[0259] Select a causal analysis algorithm:

[0260] Use appropriate causal analysis algorithms (such as Granger causality test, etc.), which are based on the lag relationship between time series data to determine whether there is a causal relationship between variables. This algorithm is suitable for analyzing the causal relationship between different factors in performance data (such as changes in task difficulty, incentive adjustment, employee skill improvement, etc.) and performance changes;

[0261] Causal Relationship Mining:

[0262] Using performance data and data on related factors that may affect performance (such as task difficulty indicators, incentive points issuance, employee training participation, etc.) as input, causal analysis operations such as Granger causality test are performed to uncover potential causes of performance changes. For example, through analysis, it is found that the sudden increase in task difficulty and the failure of employees to obtain corresponding skill training support in a timely manner are important reasons for the decline in performance scores, or that there is a significant causal relationship between newly implemented incentive measures and performance improvement, etc., to provide a basis for the subsequent formulation of management strategy adjustment plans;

[0263] Step S5-2-3: Recommend management strategy adjustment plan based on reinforcement learning:

[0264] Define the elements of the reinforcement learning environment:

[0265] Clarify the key elements of reinforcement learning, such as state space, action space, and reward function; the state space can be defined as a comprehensive state description consisting of multiple factors including current performance status (such as performance score, performance change trend, etc.), employee status (such as skill level, work attitude, etc.), and enterprise operating environment (such as market demand changes, resource allocation, etc.); the action space is the different management strategy adjustment behaviors that managers can choose (such as adjusting task allocation methods, changing incentive mechanisms, organizing training activities, etc.); the reward function is set according to the performance changes after the implementation of the management strategy. For example, if the performance is significantly improved, a higher reward value is given, otherwise a lower reward value is given, so as to measure the effectiveness of the management strategy adjustment;

[0266] Building reinforcement learning models and training:

[0267] Select appropriate reinforcement learning algorithms (such as Q-learning algorithms, deep Q network algorithms, etc.) to build reinforcement learning models. Use the historical management strategy adjustment records of the mine workshop and the corresponding performance change feedback data as training samples, so that the intelligent agent can continuously explore and learn in the simulated management decision-making environment. By repeatedly trying different management strategy adjustment actions, it updates its own strategy value function (such as Q value function) based on the reward feedback obtained, and learns the ability to select the optimal management strategy adjustment plan under different conditions.

[0268] Recommended management strategies:

[0269] When it is predicted that performance may decline or potential problems are found, based on the current actual status (the current performance status, employee status, etc. are determined by real-time collection and analysis of relevant data), the trained reinforcement learning model is used to recommend the best management strategy adjustment plan for managers. For example, specific adjustment suggestions such as "conducting special skills training for employees with poor performance, adjusting their task allocation, reducing task difficulty, and increasing corresponding incentives to encourage them to improve performance" are recommended. Managers can make corresponding management strategy adjustments in the workshop management system based on the recommended plan;

[0270] Step S5-3, feedback optimization loop:

[0271] Step S5-3-1, adjustment plan implementation and performance data feedback:

[0272] Managers adjust the plan based on the recommended management strategy and make corresponding adjustments in the actual management work of the mine workshop, such as adjusting the task allocation of employees, organizing relevant training activities, modifying the incentive mechanism, etc., to ensure that the adjustment plan can be implemented;

[0273] After the adjustment plan is implemented, continue to collect and record performance change data (such as changes in employee performance scores, improvements in task completion, etc.) and other relevant feedback information (such as employee satisfaction with the adjustment plan, changes in team collaboration atmosphere, etc.) to form a new performance data set to provide a basis for subsequent optimization analysis;

[0274] Step S5-3-2, system optimization and iterative improvement:

[0275] The new performance data collected after the implementation of the adjustment plan will be fed back to the entire performance appraisal management system. Based on this feedback information, each relevant module (such as the performance evaluation module, task difficulty adjustment module, incentive mechanism module, etc.) will use the corresponding algorithms and models (such as dynamically adjusting the weights in the multi-factor weighted model, updating the parameters of the task difficulty evaluation model, optimizing the reward standards, etc.) to perform its own optimization analysis, further improve its own functions and algorithm logic, realize the continuous iterative optimization of the entire performance appraisal management system, and continuously improve the performance appraisal management level and overall operational efficiency of the mine workshop;

[0276] Through the above detailed step S5 and its sub-step operations, the performance feedback and optimization function is fully realized. With the help of visualization and intelligent analysis technology, it not only provides employees and managers with clear and intuitive performance information and improvement suggestions, but also recommends reasonable management strategy adjustment plans through scientific prediction, causal analysis and reinforcement learning. Through the feedback optimization cycle, the effectiveness and adaptability of the entire mine workshop performance appraisal management system are continuously improved, helping the mine workshop to operate efficiently and stably.

[0277] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A mine workshop performance assessment management method, characterized by: The following steps are involved: Step S1, data collection and preprocessing: collect work data through sensor fusion, text analysis and employee self-assessment weights, and store it after multiple filling and robust standardization; Step S2, dynamic performance evaluation: Based on a multi-factor weighted model, neural networks and reinforcement learning are combined to optimize factor weights and predict employees' future performance, and the performance score output is calculated in real time; Step S3, task difficulty adjustment: using analytic hierarchy process and fuzzy comprehensive evaluation to quantify task difficulty, combined with employee capability analysis, relying on linear programming and gradient descent algorithm to adapt and optimize task difficulty in real time; Step S4, incentive mechanism: using market supply and demand and collaborative filtering algorithms to adjust the points value and customize the points plan, optimizing the reward standards based on cost-effectiveness and performance sensitivity analysis, and realizing points redemption and data feedback; Step S5, performance feedback and optimization: Generate performance reports with the help of visualization library and use text mining to generate improvement suggestions, warn of performance changes through prediction and causal analysis, and use reinforcement learning to recommend management strategy adjustment plans and provide feedback for optimization.

2. A mine workshop performance assessment management method according to claim 1, characterized in that: In the data collection and preprocessing, data collection includes: Sensors are installed on key equipment in the mine workshop, and the Kalman filter algorithm is used to fuse sensor data to obtain basic data related to equipment operation and production. At the same time, natural language processing (NLP) technology is used to analyze employee work record texts, and the key information in the texts is quantified and added to the work data. Based on the linear regression algorithm, the weight of employee self-evaluation in the overall data collection is determined according to the consistency between the employee's past self-evaluation and actual performance, and the employee work data is obtained. The employee work data is then integrated in a unified format and transmitted to the data preprocessing server with the help of big data message queue technology; among them, a linear regression model is constructed ,in, is the actual performance data set, A collection of employee self-evaluation data. is the intercept, is the regression coefficient, is the error term, and the least squares method is used to estimate and To determine the weight of employee self-assessment; Data preprocessing includes: On the data preprocessing server, the multiple imputation method MICE is used to fill in the data with missing values. Multiple prediction models are established based on the data association relationship to ensure data integrity. The robust standardization RobustScaler method is used to standardize the data. The calculation formula is: ,in, is the original data, is the standardized data, Representation data The median of Representation data The interquartile range of the preprocessed data is stored in a distributed database.

3. A mine workshop performance assessment management method according to claim 1, characterized in that: In the dynamic performance evaluation, the multi-factor weighted model construction and optimization process includes: Determine the task type, difficulty, employee historical performance, teamwork effect, and innovation results as the key factors for performance evaluation. Let these factors be , based on expert experience and historical data analysis, the weights of each factor are initially set ,and , represents the total number of key factors for performance evaluation, It is an index variable used to traverse all key factors and build a multi-factor weighted model; performance score The calculation formula is: ； Then, using the neural network algorithm, the fluctuation of employee performance and the changes in the focus of the workshop business are taken as inputs. According to the conventional training and weight adjustment mechanism of the neural network, the weights of each factor in the weighted model are adjusted in real time through back propagation. At the same time, the XGBoost algorithm is combined with reinforcement learning to build a prediction model, taking the employee's work tasks as "actions" and the performance score as "rewards". Based on the exploration learning mechanism of reinforcement learning and combined with XGBoost's ability to process historical performance data, the prediction model is optimized to accurately predict the future performance of employees. The implementation method of performance score calculation and output is as follows: input the pre-processed employee work data into the optimized multi-factor weighted model, combine it with the prediction model results, calculate the employee performance score in real time according to the established calculation logic, and output it to the subsequent modules through the API interface to provide data support for related operations.

4. A mine workshop performance assessment management method according to claim 1, characterized in that: In the task difficulty adjustment, the hierarchical analysis method includes: Construct a hierarchical model and set the target layer as the task difficulty ,The criterion layer contains relevant factors, and the judgment matrix is ​​constructed by comparing the factors in the criterion layer pairwise. ,in, Indication factors Relative to The importance of Scaling method assignment; Solving the Judgment Matrix by Eigenvector Method The largest characteristic root and its corresponding eigenvector , calculate the consistency index ,in, To determine the matrix order and check consistency, Less than average random consistency index When , the consistency is considered to be passed, and the normalized feature vector That is, the weight vector of each criterion layer factor ; The fuzzy comprehensive evaluation method includes: Set of comments , fuzzy evaluation is performed on the sub-factors under each criterion layer, and a fuzzy relationship matrix is ​​constructed. ,in, Indication factors For comments The membership degree, combined with the criterion layer factor weight vector , using the fuzzy comprehensive evaluation formula , and obtain the fuzzy comprehensive evaluation result vector of task difficulty ,in, represents the fuzzy synthesis operator, which is then converted into a specific task difficulty value through the defuzzification method. The defuzzification method is the centroid method, and the calculation formula is: ,in, for Middle elements, For comment collection Middle The quantitative value corresponding to each comment, The final task difficulty value is stored in the task management database as a reference for task difficulty adjustment; For comment collection The number of comments; Cluster analysis algorithms are used to classify employees into different ability levels based on their past performance and skill mastery. The similarities between employees are calculated and classified using the corresponding clustering algorithms. For each type of employee, the Bayesian network model is used to analyze the probability relationship between tasks of different difficulty levels and employees' positive work efficiency. The network structure and conditional probability are learned based on historical data to provide a reference for the reasonable allocation of tasks. The specific strategy for real-time adjustment of task difficulty is as follows: Based on the employee performance scores and task completion status obtained from the performance evaluation module, combined with the task difficulty evaluation model and the employee capability adaptation analysis results, the linear programming algorithm is used to calculate the task difficulty adjustment range suitable for the employee's current stage with a suitable objective function and satisfy the corresponding constraints, and the task allocation is adjusted accordingly; When employees are performing tasks, the real-time emotional state of employees obtained through wearable devices and workshop video monitoring, as well as the work efficiency tracking curve drawn using time series analysis, are combined to judge the effect of task difficulty adjustment. If it does not meet expectations, the task difficulty is quickly adjusted based on the iterative update mechanism of the gradient descent algorithm to ensure that employees maintain a positive working state under tasks of appropriate difficulty, and relevant information is promptly fed back and updated.

5. A mine workshop performance assessment management method according to claim 1, characterized in that: In the incentive mechanism, the collaborative filtering algorithm includes: Assume the user set is , the item set is , the user's rating matrix for items is , Indicates user For items Ratings of users, if the user has not rated, it is recorded as a missing value; based on the user's rating matrix ,To calculate the similarity between users, the Pearson correlation coefficient method is commonly used, and the calculation formula is ,in, and are different users, Is a user and A collection of items that have been rated. and The users are and Average rating of Indicates user and The similarity between Indicates user For items Ratings; Find target users Most similar Users form a neighbor set , predict the target user by weighted average For unrated items Rating , use collaborative filtering algorithm to analyze employee-related personalized characteristics, refer to the past performance improvement of similar employees, customize personalized points acquisition rules and distribution plans for employees according to the above algorithm, and employees will obtain points according to their performance after completing the task, which will be stored in the points account database; Adopt a simulation algorithm based on market supply and demand theory, and dynamically adjust the corresponding points value according to the basic principles of the algorithm according to the popularity of different rewards among employees and the popularity of redemption; Dynamic adjustments to reward standards include: From the perspective of enterprise cost-effectiveness, the cost-effectiveness analysis model is used in combination with the dynamic programming algorithm. Based on the enterprise's business objectives and financial status, the reward standards are optimized in real time according to the corresponding algorithm logic. The difference analysis and elasticity analysis statistical methods are used to analyze the sensitivity between performance scores and changes in reward standards, so as to fine-tune the reward standards corresponding to different performance intervals. Employees choose to redeem rewards on the reward redemption platform based on their points and standards. The incentive mechanism module feeds back the redemption records to other related modules for subsequent optimization analysis. By comprehensively considering the costs and benefits of rewards, analyzing the relationship between performance and rewards, setting reward standards scientifically and reasonably, employees are encouraged to improve their performance, and at the same time, the incentive mechanism and the entire performance appraisal management system are coordinated and optimized.

6. A mine workshop performance assessment management method according to claim 1, characterized in that: In the performance feedback and optimization, management strategy adjustment and optimization include: Use time series prediction algorithms to predict trends in team and individual performance data, and combine causal analysis algorithms to explore potential causes of performance changes. When the probability of performance decline or potential problems is predicted, the early warning mechanism is triggered according to the set threshold and the early warning information is pushed to workshop managers. Based on the reinforcement learning algorithm, by simulating the performance changes after the implementation of different management strategies, the optimal management strategy adjustment plan is recommended to managers according to the reinforcement learning strategy recommendation mechanism. Managers make corresponding adjustments in the workshop management system accordingly. The adjusted performance change data is fed back to the system for further optimization of the parameters and models of each algorithm module, so as to realize continuous iterative optimization of the entire performance appraisal management system and improve the overall operational efficiency of the mine workshop.

Images