Dynamic performance appraisal management system and method based on full-process digitization

By using a fully digital dynamic performance appraisal management system, natural language processing and cosine similarity algorithms are used to dynamically allocate indicator weights. Combined with cross-system spatiotemporal alignment and abnormal data compensation mechanisms, the static nature and insufficient environmental adaptability of enterprise performance appraisal systems are solved, achieving high-precision and comprehensive appraisal results.

CN120912064APending Publication Date: 2025-11-07CCCC SOUTH CHINA SURVEY & MAPPING TECH CO LTD +1
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202511112025.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-08-08
Publication Date
2025-11-07

AI Technical Summary

Technical Problem

Existing corporate performance appraisal systems suffer from staticity, lag, and insufficient environmental adaptability, making it difficult to quickly respond to strategic adjustments and multi-source data fusion analysis, resulting in one-sided and unscientific appraisal results.

Method used

The dynamic performance appraisal management system adopts a fully digital process, dynamically allocating indicator weights through natural language processing and cosine similarity algorithm, combined with cross-system spatiotemporal alignment and abnormal data compensation mechanism, monitoring corporate environmental variables to adjust appraisal standards, building an improvement suggestion knowledge base and using blockchain for evidence storage.

Benefits of technology

It has enabled dynamic response to strategic iterations, improved the environmental adaptability and data accuracy of the assessment system, and enhanced the objectivity and comprehensiveness of the assessment results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120912064A_ABST
    Figure CN120912064A_ABST
Patent Text Reader

Abstract

The invention discloses a dynamic performance appraisal management system and method based on full-process digitization. The method comprises the following steps: collecting working behaviors, business achievements and environmental data of employees through dynamic authorization; after preprocessing, extracting execution efficiency, cooperative influence and environment adaptation type characteristic indexes; an enterprise strategy target keyword and an employee work target keyword in the enterprise strategy document and the employee data are extracted by using a natural language processing technology, and the matching degree of an employee target and an enterprise strategy target is calculated in combination with a cosine similarity algorithm; dynamically distributing strategic weights of the characteristic indexes; setting an assessment standard threshold value, and monitoring an enterprise business environment variable to dynamically adjust the assessment standard threshold value; and calculating a comprehensive score of the employee and generating an improvement suggestion in combination with the feature indexes, the strategic weights and the assessment standard threshold. According to the invention, through full-process digital management of strategic target dynamic adaptation, cross-system data accurate alignment and environment self-adaptation, the assessment objectivity and the business response capability are significantly improved.
Need to check novelty before this filing date? Find Prior Art

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of enterprise performance management, in particular to a dynamic performance evaluation management system and method based on full-process digitization. BACKGROUND

[0002] The current enterprise performance evaluation system generally has the problems of static, hysteresis and insufficient environmental adaptability, which seriously restricts the dynamic landing of strategic goals and the accurate evaluation of business execution. The traditional performance evaluation method mainly adopts a fixed weight distribution mechanism, which relies on manual presetting of index weights and is difficult to quickly respond to the needs of enterprise strategic adjustment. When the strategic goal is significantly diverted due to market changes, the original weight system cannot automatically associate new strategic elements, and needs to be reconfigured through complex manual intervention, resulting in a significant time lag between evaluation orientation and strategic execution. In addition, the existing system is limited to a single business data dimension, and lacks multi-source fusion analysis of employee behavior data and environmental parameters, resulting in one-sided evaluation results and difficulty in fully reflecting the real contribution of employees in complex scenarios.

[0003] In the prior art, there are obvious shortcomings in data collection and processing. Most systems rely on dispersed enterprise information systems to obtain data, but cross-system data is difficult to effectively integrate due to problems such as heterogeneous formats, time sequence misplacement and missing entity association, resulting in insufficient feature index extraction accuracy. The behavior data of employees on the collaborative office platform and the business results data of the ERP system cannot accurately establish causal relationships due to the lack of unified spatio-temporal alignment rules, which affects the scientificity of evaluation indicators. At the same time, traditional abnormal data processing mainly uses simple threshold filtering or mean filling, without considering the specificity of business scenarios, which may cause the deletion of effective information or the residual of noise interference. SUMMARY

[0004] In order to solve the technical problems mentioned in the background art, the purpose of the present application is to provide a dynamic performance evaluation management system and method based on full-process digitization.

[0005] In order to achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0006] A dynamic performance evaluation management system based on full-process digitization, comprising:

[0007] An employee data management module: collecting first employee data within the authorized scope of employees; preprocessing the first employee data to obtain second employee data, and extracting quantifiable feature indicators based on the second employee data;

[0008] Strategy dynamic adaptation module: extracting enterprise strategy target keywords and employee work target keywords in enterprise strategy documents and the second employee data through natural language processing technology, calculating the matching degree of employee targets and enterprise strategy targets through cosine similarity algorithm, and dynamically assigning the strategic weight of the feature index according to the matching degree;

[0009] Intelligent performance evaluation module: setting an evaluation standard threshold, and monitoring enterprise business environment variables to dynamically adjust the evaluation standard threshold; combining the feature index, the strategic weight and the evaluation standard threshold to calculate the comprehensive score of the employee;

[0010] Improvement and evidence storage module: constructing an improvement suggestion knowledge base to automatically match improvement suggestion texts for employees; using blockchain technology to store the second employee data and the comprehensive score.

[0011] Further, the first employee data includes but is not limited to employee work behavior, business results, work environment and employee comments, and the first employee data comes from different enterprise systems.

[0012] Further, the preprocessing includes data alignment and association, data compensation and rejection, data format standardization and desensitization processing;

[0013] The data alignment and association refers to the temporal and spatial alignment of the first employee data of the same employee in different enterprise systems and the data association with each employee as an entity, and each employee is given a unique identifier;

[0014] The data compensation and rejection refers to interpolating compensation for null values in the first employee data and rejecting abnormal data;

[0015] The data format standardization refers to unifying the format of the first employee data from different enterprise systems;

[0016] The desensitization processing refers to irreversible hash replacement of sensitive information in the first employee data.

[0017] Further, the feature index includes different types of feature indexes such as execution efficiency index, synergistic influence index and environmental adaptation index.

[0018] Further, the enterprise strategy target keywords include department-level keywords and post-level keywords as two-level strategy indexes;

[0019] Scanning enterprise strategy documents again, respectively converting strategy target keywords contained in two-level strategy indexes into vector encoding:

[0020]

[0021] Wherein, is the element of the jth employee work goal vector, d is the frequency of the ith department-level keyword appearing in the enterprise strategy document, is the average frequency of the ith department-level keyword appearing in the industry standard document, is the pre-calculated importance of the ith department-level keyword, is the preset weight of the ith department-level keyword;

[0022] wherein, is the element of the jth employee work goal vector, p is the frequency of the Kth post-level keyword appearing in the enterprise strategy document, is the average frequency of the Kth post-level keyword appearing in the industry standard document, is the pre-calculated importance of the Kth post-level keyword, is the preset weight of the Kth post-level keyword;

[0023] scanning the second employee data again, converting the employee work goal keywords contained in the second employee data into vector encoding:

[0024]

[0025] wherein, E j is the element of the jth employee work goal vector, j is the value coefficient of the jth employee work goal keyword, is the semantic included angle between the jth employee work goal keyword and the department-level strategic indicator vector S d is the semantic included angle between the jth employee work goal keyword and the post-level strategic indicator vector S p , and ω j is the preset weight of the jth employee work goal keyword;

[0026] using the cosine similarity algorithm to calculate the matching degree between the enterprise strategy goal and the employee work goal:

[0027]

[0028] wherein, S d is the set of department-level strategic indicator vector elements, S p is the set of post-level strategic indicator vector elements, E is the set of vector elements of all employee work goal keywords, E = {E1, E2, E3...E j ​​​λ is a preset department-level weight.

[0029] Further, the high matching degree employee, the medium matching degree employee and the low matching degree employee are divided according to the matching degree;

[0030] The strategic weight of the execution efficiency type index of the high matching degree employee is improved;

[0031] The strategic weight of the coordination influence type index of the medium matching degree employee is improved;

[0032] The strategic weight of the environment adaptation type index of the low matching degree employee is improved.

[0033] Further, the enterprise business environment variables are divided into five types of policy and regulation, market fluctuation, technical trend, internal operation and social environment, and the influence weights of the five environment types on the enterprise business execution are set respectively according to the influence degree;

[0034] Based on the environment variables, the evaluation standard threshold is calculated, and the calculation formula of the evaluation standard threshold is as follows:

[0035]

[0036] Wherein, T new is the adjusted evaluation standard threshold, T base is a preset evaluation standard threshold, β is a preset environment sensitivity coefficient, ΔH is an actual change amount of the environment variable, H ref is a reference value of the environment variable;

[0037] The actual change amount of the environment variable is calculated according to the following formula:

[0038]

[0039] Wherein, is the actual observation value of the yth environment variable at the monitoring time point, is the historical normal reference value of the yth environment variable, H (ref) is the preset environment variable tolerance threshold of the enterprise, F y is the influence weight of the yth environment variable, τ is a time attenuation coefficient, t y is a time delay factor.

[0040] Further, the comprehensive score of the employee is calculated according to the following formula:

[0041]

[0042] Wherein, R is the comprehensive score of the employee, w u is the strategic weight of the u th feature index, x v is the value of the v th feature index, nu Total number of the u-th type characteristic index, R u Score R of the u-th type characteristic index u .

[0043] Further, the improved suggestion knowledge base stores historical improvement suggestion texts and implementation effects of the improvement suggestions.

[0044] The score of the characteristic index and the employee comprehensive score are input into the improvement knowledge base to automatically match the historical improvement suggestion texts.

[0045] A dynamic performance assessment management method based on full-process digitization, the method comprising:

[0046] Step 1: collecting first employee data within the authorized scope of an employee; preprocessing the first employee data to obtain second employee data, and extracting a quantifiable characteristic index based on the second employee data;

[0047] Step 2: extracting enterprise strategic target keywords and employee work target keywords from enterprise strategic documents and the second employee data through natural language processing technology, calculating the matching degree of employee targets and enterprise strategic targets through a cosine similarity algorithm, and dynamically assigning strategic weights of the characteristic index according to the matching degree;

[0048] Step 3: setting an assessment standard threshold and monitoring enterprise business environment variables to dynamically adjust the assessment standard threshold; combining the characteristic index, the strategic weight and the assessment standard threshold to calculate a comprehensive score of the employee;

[0049] Step 4: constructing an improvement suggestion knowledge base, automatically matching improvement suggestion texts for the employee; and using blockchain technology to store the second employee data and the comprehensive score.

[0050] Compared with the prior art, the present application has the following advantages:

[0051] 1. The present application analyzes enterprise strategic targets through natural language processing, combines a cosine similarity algorithm to quantitatively calculate the matching degree of employee targets and strategies in real time, and dynamically assigns index weights, thereby solving the lag problem of the traditional fixed weight mechanism that cannot respond to strategic iteration.

[0052] 2. The present application innovates a space-time alignment rule and an abnormal data compensation mechanism, fuses multi-source heterogeneous data, and guarantees data security while improving the extraction accuracy of characteristic indexes through unique identifiers and sensitive information desensitization technology.

[0053] 3. Based on business environment variable monitoring, the assessment standard is adjusted in real time, and compared with the traditional static threshold system, the adaptability of the assessment system to external environment variables is significantly enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0055] Figure 1 The system workflow schematic diagram of the present application;

[0056] Figure 2 The employee data acquisition schematic diagram of the present application;

[0057] Figure 3 The process schematic diagram of the strategic weight distribution of the feature index of the present application. DETAILED DESCRIPTION

[0058] To achieve the above purpose, the present application is implemented by the following technical solutions, and the present application provides a dynamic performance evaluation management system based on full-process digitization, please refer to Figure 1 The system comprises:

[0059] The employee data management module: collecting the first employee data within the authorized range of the employee; pre-processing the first employee data to obtain the second employee data, and extracting the quantifiable feature index based on the second employee data;

[0060] The employee authorization adopts dynamic on-demand authorization, and the system automatically adjusts the collection permission of the first employee data according to the current work task and work scene of the employee, and at the same time, the employee can also manually modify the collection permission of the first employee data;

[0061] The first employee data includes but is not limited to the work behavior, business results, work environment and employee comments of the employee, and the first employee data comes from different enterprise systems, please refer to Figure 2 ;

[0062] The personnel behavior acquisition is realized by the following technical means:

[0063] 1) Through the lightweight acquisition agent program deployed on the office terminal, the behavior logs such as attendance punch-in, system login, document operation, etc. are captured in real time;

[0064] 2) The conference room intelligent terminal automatically associates the participants through voiceprint recognition technology and generates meeting contribution data.

[0065] The business results are obtained by interfacing with various business systems of the enterprise:

[0066] 1) The project management system provides engineering data such as task completion degree, milestone achievement rate, etc.,

[0067] 2) Customer Relationship Management System outputs customer satisfaction, opportunity conversion rate, etc.

[0068] 3) Enterprise Resource Planning System obtains cost control, revenue contribution, etc. Economic data, collect email confirmation letter, acceptance report and other document information.

[0069] Work environment collection collects physical work environment parameters through Internet of Things sensors, including temperature and humidity, equipment utilization, and production environment data.

[0070] Employee comments are obtained through self-evaluation and mutual evaluation of employees in the system.

[0071] The first employee data preprocessing includes data alignment and association, data compensation and rejection, data format standardization and desensitization processing.

[0072] Data alignment and association refer to the temporal and spatial alignment of the first employee data of the same employee in different enterprise systems and the data association of each employee as an entity, and each employee is assigned a unique identifier. Temporal and spatial alignment includes time dimension alignment and space dimension alignment.

[0073] In this embodiment, time dimension alignment is achieved by marking each system data with a unified clock source, reordering the behavior logs of different sources (such as attendance punch-in, meeting records) according to the actual occurrence time, and for the data that arrives with delay, a sliding time window mechanism is used for dynamic collection, ensuring the logical coherence of the operation time sequence across systems,

[0074] Time dimension alignment formula:

[0075]

[0076] Where A(t b ) is the window first employee data at time t b , ΔT1 is the sliding window length (default 15 minutes), D(τ) is the data volume of the first employee data at time τ, is the decay coefficient (default 0.05 / s), and t-τ is the data delay time difference.

[0077] In this embodiment, space dimension alignment is achieved by establishing a mapping matrix of department-post-physical location, converting the location information (such as workstation number, conference room ID) scattered in multiple enterprise systems into a unified virtual coordinate system.

[0078] In this embodiment, data association is achieved by automatically associating the scattered records of the same employee in different enterprise systems through employee ID, enterprise mailbox, biometric features (such as voiceprint template), etc. For records with information conflicts (such as differences in name spelling in different systems), an artificial review process is started.

[0079] Data compensation and rejection refers to interpolating compensation for null values in the first employee data and rejecting abnormal data;

[0080] In this embodiment, the system monitors the first employee data in real time, identifies missing values of employee behavior, business results and other fields, compensates for missing time series data (such as attendance records) by using a sliding time window to analyze the rules of adjacent periods, and compensates for missing non-time series data (such as department performance indicators) based on historical data of personnel in the same position to ensure data continuity.

[0081] The linear interpolation compensation formula for time series data is as follows:

[0082]

[0083] Wherein, R t is the interpolation result at time t, and are the original data of adjacent effective times t1 and t2, and t is the time of the current missing data;

[0084] The compensation formula for non-time series data is as follows:

[0085]

[0086] Wherein, is the missing value compensation result of department d, is the historical data of the zth employee in the same position in department d, and Z is the total number of employees in department d.

[0087] In this embodiment, the abnormal data detection uses standard deviation analysis to identify extreme values deviating by more than 3σ;

[0088] The standard deviation calculation formula is as follows:

[0089]

[0090] Wherein, μ is the average value of the current first employee data, is the total number of first employee data, Q a is the a th observation value in the data set,

[0091] The deviation calculation formula is as follows:

[0092]

[0093] Wherein, P is the deviation, and G is the data point value to be detected;

[0094] When the deviation of the first employee data |P|>3, it is determined as abnormal data and is rejected, otherwise it is normal first employee data.

[0095] Data format standardization, unifying the format of the first employee data from different enterprise systems,

[0096] In this embodiment, data format standardization includes data unified modeling, heterogeneous data format conversion and semantic consistency alignment.

[0097] Data unified modeling defines a global data dictionary to clearly define the standard name, data type and format specification of each data field, for example: time class data uniformly uses ISO 8601 standard and numerical data specifies decimal place accuracy (such as cost data retains 2 decimal places) and unit system (such as currency is unified to RMB yuan) etc.

[0098] Heterogeneous data format conversion implements adaptation to the original data characteristics of different enterprise systems by building a format conversion rule library, for example, text cleaning removes special symbols, full and half angle conversion (such as converting "2023 year" to "2023 year") and encoding unified character set is forced to convert to UTF-8, eliminating the risk of garbled code, etc.

[0099] Semantic consistency alignment eliminates the differentiated expression of business concepts through semantic analysis, for example: synonym normalization unifies the expressions of "KPI", "performance indicator", "evaluation item" and other systems to "feature indicator", and the measurement standard calibration establishes a linear conversion formula for indicators with calculation differences (such as 1-5 point system and percentage system of "customer satisfaction").

[0100] Sensitive information desensitization processing refers to irreversible hash replacement of sensitive information of employees,

[0101] In this embodiment, the sensitive fields are converted using an irreversible hash function to map the original data to a fixed-length unique hash value, and this process enhances security by adding a system-generated unique random string (salt value) to prevent reverse cracking through rainbow tables, etc.

[0102] In this embodiment, the feature indicators include different types of feature indicators such as performance efficiency indicators, collaborative influence indicators and environmental adaptation indicators through the second employee data extraction.

[0103] Strategy dynamic adaptation module: extracts enterprise strategy target keywords and employee work target keywords from enterprise strategy documents and the second employee data through natural language processing technology, calculates the matching degree of employee targets and enterprise strategy targets through cosine similarity algorithm, and dynamically allocates the strategic weight of the feature indicators according to the matching degree;

[0104] The allocation steps of the strategic weight are as follows, please refer to Figure 3 :

[0105] The enterprise strategic target keywords include department-level keywords and post-level keywords as two levels of strategic indicators;

[0106] By scanning the enterprise strategic document again, the strategic target keywords contained in the two levels of strategic indicators are respectively converted into vector encoding:

[0107]

[0108] Wherein, is the element of the a-th component department-level strategic indicator vector S d , is the frequency of the i-th department-level keyword appearing in the enterprise strategic document, is the average frequency of the i-th department-level keyword appearing in the industry standard document (such as industry white paper, national standard document and market research report), is the pre-calculated importance of the i-th department-level keyword, is the preset weight of the i-th department-level keyword;

[0109] Wherein, is the element of the o-th component post-level strategic indicator vector S p , is the frequency of the K-th post-level keyword appearing in the enterprise strategic document, is the average frequency of the K-th post-level keyword appearing in the industry standard document (such as industry white paper, national standard document and market research report), is the pre-calculated importance of the K-th post-level keyword, is the preset weight of the K-th post-level keyword;

[0110] In this embodiment, the calculation formula of the importance of the department-level keyword is as follows:

[0111]

[0112] Wherein, B i is the semantic weight of the i-th department-level keyword in the enterprise strategic document, C i is the contribution degree of the i-th department-level keyword in the enterprise strategic document in business results, τ is a time decay coefficient, t a is the number of years from the publication time of the enterprise strategic document, g、 and q are preset weight coefficients.

[0113] In this embodiment, the calculation formula of the importance of the post-level keyword is as follows:

[0114]

[0115] Wherein, B KC represents the semantic weight of the Kth job-level keyword in the corporate strategy document. K t represents the contribution of the Kth job-level keyword in the corporate strategy document to business results. a The number of years since the release of the company's strategic documents, g. q are preset weighting coefficients.

[0116] In this embodiment, semantic weights are obtained by the position of department-level keywords and job-level keywords in the corporate strategy document. For example, higher weights are given when department-level keywords and job-level keywords appear in the title, the first paragraph of a chapter, or other positions in the strategy document.

[0117] Scan the second employee data again and convert the employee work goal keywords in the second employee data into vector codes:

[0118]

[0119] Among them, E j Let α be the element of the work goal vector for the j-th employee. j The value coefficient of the keywords for the work goals of the j-th employee. For the j-th employee's work goal keywords and departmental strategic indicator vector S d The semantic angle, The vector S represents the keyword of the j-th employee's work objectives and the strategic indicator at the job level. p The semantic angle, ω j The preset weights for the keywords related to the work goals of the j-th employee;

[0120] In this embodiment, the value coefficient of employee work goal keywords is a quantitative parameter for the contribution value of employee work goal keywords to business. It is used to measure the expected impact of employee work goal keywords on the actual business output of the enterprise and is obtained by the enterprise itself.

[0121] The cosine similarity algorithm is used to calculate the degree of alignment between corporate strategic goals and employee work goals:

[0122]

[0123] Among them, S d This is a set of vector elements representing departmental-level strategic indicators. S p This is a set of vector elements representing job-level strategic indicators. E is the set of vector elements representing the keywords of all employees' work goals, E = {E1, E2, E3, ..., E...} j}; λ is the preset department-level weight.

[0124] Strategic weights for feature metrics are automatically assigned based on matching degree, specifically including:

[0125] According to the matching degree, the employees are divided into high matching degree employees, medium matching degree employees and low matching degree employees;

[0126] The strategic weight of the execution efficiency type index of the high matching degree employee is improved;

[0127] The strategic weight of the coordination influence type index of the medium matching degree employee is improved;

[0128] The strategic weight of the environment adaptation type index of the low matching degree employee is improved.

[0129] The intelligent performance evaluation module: set the evaluation standard threshold, and monitor the enterprise business environment variables to dynamically adjust the evaluation standard threshold; combine the characteristic index, the strategic weight and the evaluation standard threshold to calculate the comprehensive score of the employee;

[0130] In this embodiment, monitoring the enterprise business environment variables is achieved by connecting the API interface of public data and purchasing third-party commercial data services (such as Tianyancha and SimilarWeb).

[0131] The enterprise business environment variables are divided into five types: policies and regulations, market fluctuations, technical trends, internal operations and social environment, and the influence weight of each type of environment is set according to the influence degree of the enterprise business execution;

[0132] Based on the environment variables, the evaluation standard threshold is calculated, and the calculation formula of the evaluation standard threshold is as follows:

[0133]

[0134] Wherein, T new is the adjusted evaluation standard threshold, T base is the preset evaluation standard threshold, β is the preset environment sensitivity coefficient, ΔH is the actual change amount of the environment variable, H ref is the reference value of the environment variable;

[0135] The actual change amount of the environment variable is calculated according to the following formula:

[0136]

[0137] Wherein, is the actual observation value of the yth type of environment variable at the monitoring time point, is the historical normal reference value of the yth type of environment variable, H (ref) is the preset environment variable tolerance threshold of the enterprise, F y is the influence weight of the yth type of environment variable, τ is the time attenuation coefficient, t yis a time delay factor, the time delay factor represents a delay interval between a time when an actual change of an environment variable of the yth type (such as a policy and regulation, market fluctuation, etc.) occurs and a time when the system monitors the change.

[0138] The comprehensive score of the employee is calculated by the following formula:

[0139]

[0140] wherein R is the comprehensive score of the employee, w u is a strategic weight of the u th characteristic index, x v is a value of the v th characteristic index, n u is a total number of the u th index, R u is a score R of the u th characteristic index, u .

[0141] The improvement and evidence storage module: constructing an improvement suggestion knowledge base to automatically match improvement suggestion texts for the employee; using blockchain technology to store the second employee data and the comprehensive score;

[0142] The improvement suggestion knowledge base stores historical improvement suggestion texts and implementation effects of the improvement suggestions;

[0143] The historical improvement suggestion texts are automatically matched by inputting the scores of the characteristic indexes and the comprehensive score of the employee into the improvement knowledge base.

[0144] A dynamic performance assessment management method based on full-process digitization, the method comprising:

[0145] Step 1: collecting first employee data within the authorized scope of the employee; pre-processing the first employee data to obtain second employee data, and extracting characteristic indexes that can be quantitatively calculated based on the second employee data;

[0146] Step 2: extracting enterprise strategic target keywords and employee work target keywords from enterprise strategic documents and the second employee data by using natural language processing technology, calculating the matching degree of the employee target and the enterprise strategic target by using a cosine similarity algorithm, and dynamically assigning strategic weights of the characteristic indexes according to the matching degree;

[0147] Step 3: setting an assessment standard threshold, and monitoring enterprise business environment variables to dynamically adjust the assessment standard threshold; combining the characteristic indexes, the strategic weights and the assessment standard threshold to calculate the comprehensive score of the employee;

[0148] Step 4: constructing an improvement suggestion knowledge base to automatically match improvement suggestion texts for the employee; using blockchain technology to store the second employee data and the comprehensive score.

[0149] In summary, the application has the advantages of innovatively constructing a cross-system space-time alignment rule and an intelligent compensation mechanism for abnormal data, breaking through the barriers of multi-modal data scattered in behavior logs, business systems and environmental sensors, and realizing high-precision feature index extraction under the premise of ensuring data security through sensitive information desensitization, semantic consistency calibration and other technologies, thereby enhancing the objectivity and comprehensiveness of the assessment results.

[0150] Finally: the above only for the preferred embodiments of the present application, and not for limiting the present application, any modification within the spirit and principles of the present application, equivalent replacement, improvement, etc., should be included in the protection scope of the present application.

Claims

1. A dynamic performance appraisal management system based on full-process digitization, characterized in that, The method comprises the following steps: An employee data management module: collects first employee data within the scope of employee authorization; Preprocessing the first employee data to obtain second employee data, and extracting quantifiable feature indicators based on the second employee data; A strategic dynamic adaptation module: extracts enterprise strategic target keywords and employee work target keywords from enterprise strategic documents and the second employee data through natural language processing technology, calculates the matching degree of employee targets and enterprise strategic targets through a cosine similarity algorithm, and dynamically allocates strategic weights to the feature indicators according to the matching degree; An intelligent performance evaluation module: sets an evaluation standard threshold, and monitors enterprise business environment variables to dynamically adjust the evaluation standard threshold; combines the feature indicators, the strategic weights, and the evaluation standard threshold to calculate the comprehensive score of the employee; An improvement and evidence storage module: constructs an improvement suggestion knowledge base to automatically match improvement suggestion texts for employees; and stores the second employee data and the comprehensive score by using a blockchain technology.

2. The system of claim 1, wherein, The first employee data includes but is not limited to employee work behavior, business results, work environment, and employee comments, and the first employee data comes from different enterprise systems.

3. The system of claim 2, wherein, The preprocessing includes data alignment and association, data compensation and rejection, data format standardization, and desensitization processing; The data alignment and association refers to the temporal and spatial alignment of the first employee data of the same employee in different enterprise systems, and the data association of each employee as an entity, and giving each employee a unique identifier; The data compensation and rejection refers to interpolating compensation for null values in the first employee data and rejecting abnormal data; The data format standardization refers to unifying the format of the first employee data from different enterprise systems; The desensitization processing refers to irreversible hash replacement of sensitive information in the first employee data.

4. The system of claim 3, wherein, The feature indicators include different types of feature indicators such as execution efficiency indicators, synergistic influence indicators, and environment adaptation indicators.

5. The system of claim 4, wherein, The enterprise strategic target keywords include department-level keywords and post-level keywords as two-level strategic indicators; The enterprise strategic documents are scanned again, and the strategic target keywords contained in the two-level strategic indicators are converted into vector encoding respectively: wherein, S is the vector of strategic indicators of the a-th component department level strategy, d is the element of S, is the frequency of occurrence of the i-th department level keyword in the enterprise strategy document, is the average frequency of occurrence of the i-th department level keyword in the industry standard document, is the pre-computed importance of the i-th department level keyword, is the pre-set weight of the i-th department level keyword. wherein, S is the o-th component post-level strategic indicator vector, p is the element of the o-th component post-level strategic indicator vector S, is the frequency of the K-th post-level keyword in the enterprise strategy document, is the average frequency of the K-th post-level keyword in the industry standard document, is the pre-computed importance of the K-th post-level keyword, is the pre-set weight of the K-th post-level keyword; The second employee data is scanned again, and the employee work target keywords contained in the second employee data are converted into vector encoding: wherein, E j is the jth employee work goal vector element, α j is the value coefficient of the jth employee work goal keyword, is the semantic included angle between the jth employee work goal keyword and the department-level strategic indicator vector S d , ω p is the semantic included angle between the jth employee work goal keyword and the post-level strategic indicator vector S j , and ω d is the preset weight of the jth employee work goal keyword. The cosine similarity algorithm is used to calculate the matching degree of the enterprise strategic targets and the employee work targets: Among them, S d This is a set of vector elements representing departmental-level strategic indicators. S p This is a set of vector elements representing job-level strategic indicators. E is the set of vector elements representing the keywords of all employees' work goals, E = {E1, E2, E3, ..., E...} j }; λ is the preset department-level weight.

6. The system of claim 5, wherein, According to the matching degree, high-matching-degree employees, medium-matching-degree employees, and low-matching-degree employees are divided; The strategic weight of the execution efficiency indicator of the high-matching-degree employee is increased; The strategic weight of the synergistic influence indicator of the medium-matching-degree employee is increased; The strategic weight of the environment adaptation indicator of the low-matching-degree employee is increased.

7. The system of claim 6, wherein, The enterprise business environment variables are divided into five types: policies and regulations, market fluctuations, technology trends, internal operations, and social environment, and the influence weights of the five environment types on enterprise business execution are set respectively; Based on the environment variables, the evaluation standard threshold is calculated, and the calculation formula of the evaluation standard threshold is as follows: Wherein, T new is the adjusted evaluation standard threshold, T base is the preset evaluation standard threshold, β is the preset environment sensitivity coefficient, ΔH is the actual change of the environment variable, H ref is the environment variable reference value; The actual change amount of the environment variable is calculated according to the following formula: wherein, is the actual observed value of the yth environmental variable at the monitoring time point, is the historical normal reference value of the yth environmental variable, H (ref) is the pre-set environmental variable tolerance threshold of the enterprise, F y is the impact weight of the yth environmental variable, τ is the time decay coefficient, t y is the time delay factor.

8. The system of claim 7, wherein, The comprehensive score calculation formula of the employee is: where R is the overall score of the employee, w u is the strategic weight of the u-th characteristic indicator, x v is the value of the v-th characteristic indicator, n u is the total number of characteristic indicators of the u-th type, R u is the score R of the u-th characteristic indicator u .

9. The system of claim 8, wherein, The improvement suggestion knowledge base stores historical improvement suggestion texts and implementation effects of the improvement suggestions; The historical improvement suggestion texts are automatically matched by inputting the score of the characteristic index and the comprehensive score of the employee into the improvement knowledge base.

10. A dynamic performance appraisal management method based on full-process digitization, characterized in that, The method comprises: Step 1: collecting first employee data within the authorized scope of the employee; pre-processing the first employee data to obtain second employee data, and extracting a quantifiable characteristic index based on the second employee data; Step 2: extracting enterprise strategic target keywords and employee work target keywords from enterprise strategic documents and the second employee data through natural language processing technology, calculating the matching degree of the employee target and the enterprise strategic target through a cosine similarity algorithm, and dynamically assigning a strategic weight of the characteristic index according to the matching degree; Step 3: setting an evaluation standard threshold, and monitoring enterprise business environment variables to dynamically adjust the evaluation standard threshold; combining the characteristic index, the strategic weight and the evaluation standard threshold to calculate the comprehensive score of the employee; Step 4: constructing an improvement suggestion knowledge base, automatically matching improvement suggestion texts for the employee; and using blockchain technology to store the second employee data and the comprehensive score.

Citation Information

Patent Citations

  • Comprehensive performance management method and system based on strategic goals

    CN115222259A

  • Performance management method and device based on enterprise strategic target and intelligent platform

    CN117291465A

  • Enterprise employee dynamic visualization performance assessment automatic evaluation system

    CN119624256A