A talent evaluation report generation method and system

Abstract

The present application relates to the technical field of talent management, and specifically discloses a talent evaluation report generation method and system, which first initializes an evaluation scene and a post image, and determines an evaluation dimension set matching post requirements; then collects multi-source heterogeneous data through internal and external multi-channels and performs standardized processing to ensure data security compliance; adopts a way of combining analytic hierarchy process and reinforcement learning to dynamically allocate weights, quantifies and qualitatively scores each index, and calculates a comprehensive score; improves evaluation accuracy through model self-iteration optimization, and finally generates a personalized evaluation report containing multiple modules and supports multi-format output and batch generation. The present application solves the problems of traditional evaluation standard solidification and data fragmentation, improves evaluation pertinence and efficiency, helps precise talent management, and has strong practicality.

Description

Technical Field

[0001] This invention relates to the field of talent management technology, specifically to a method and system for generating talent evaluation reports. Background Technology

[0002] In today's increasingly competitive talent environment, organizations have a growing need for scientific talent evaluation. Talent evaluation reports, as the core basis for talent management decisions, directly impact the quality of talent management work due to their efficiency, accuracy, and comprehensiveness. Existing talent evaluation methods suffer from the following shortcomings: 1. The evaluation criteria are rigid and singular, often adopting a crude model of "education + years of service" and "KPI + subjective scoring". This model has a low degree of matching with the actual needs of the job and makes it difficult to take into account the individual characteristics and dynamic growth trajectory of talents. 78% of enterprises are still troubled by this problem, resulting in limited reference value of the evaluation results. 2. Data collection is fragmented, with data from multiple sources such as HR systems, OA systems, and training platforms being disconnected from each other, resulting in incomplete talent profiles, weak data support, and an inability to fully capture multi-dimensional information such as talent's abilities, performance, and potential; 3. The evaluation process is cumbersome, relying heavily on manual data entry, review, and analysis, which consumes a lot of manpower and time. Furthermore, human intervention is prone to introducing subjective bias, resulting in low evaluation efficiency and unfair results. 4. The report generation lacks flexibility and cannot dynamically adjust the evaluation dimensions and weights according to different evaluation scenarios and job types. The generated reports are highly homogeneous and cannot meet the needs of personalized decision-making. 5. The evaluation results lack interpretability and dynamic iteration capability, fail to clearly present the correlation between evaluation indicators and evaluation results, and cannot optimize the evaluation model in real time based on changes in talent growth data and business needs, making it difficult for the evaluation system to adapt to organizational development.

[0003] Currently, existing technologies have significant shortcomings in multi-source data fusion, dynamic weight adjustment, personalized report generation, and self-iterative evaluation models. Therefore, there is an urgent need for a new method and system for generating talent evaluation reports to achieve intelligent, accurate, and personalized talent evaluation throughout the entire process, providing reliable support for talent management decisions. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for generating talent evaluation reports, which solves the problems of rigid talent evaluation standards, fragmented data, cumbersome processes, homogenized reports, and models that cannot self-iterate. It achieves deep integration of multi-source data, dynamic adaptation of evaluation dimensions, intelligent and efficient evaluation process, and personalized generation of evaluation reports. At the same time, it has the ability to self-iterate the evaluation model, improves the accuracy and practicality of talent evaluation, and provides a scientific, comprehensive, and reliable basis for organizational talent management decisions.

[0005] To solve the above-mentioned technical problems, the technical solution provided by the present invention is: a method for generating talent evaluation reports, comprising the following steps: Step S1: Initialize the evaluation scenario and job profile S1.1 Receives evaluation scenario parameters input by the user, the evaluation scenario parameters including evaluation purpose, evaluation period, and evaluation object scope; S1.2 Based on the preset job classification library, obtain the job type corresponding to the evaluation object, call the job profile generation module, and generate a job competency profile by combining the organization's business needs and the core responsibilities of the job. The job competency profile includes core competency dimensions, auxiliary competency dimensions, competency requirement thresholds, and priority of each dimension. The core competency dimension is dynamically determined according to the job type. S1.3 Based on the evaluation scenario parameters and job competency profile, initialize the evaluation dimension set. The evaluation dimension set includes basic information dimension, competency dimension, performance dimension, potential dimension, and behavior dimension. Each dimension contains several specific evaluation indicators, and the selection of indicators is precisely matched with the job profile.

[0006] Step S2: Multi-source heterogeneous data acquisition and standardization processing S2.1 Construct a multi-source data acquisition channel, which includes an internal data channel and an external data channel: Internal data channel: Connects to the organization's internal HR system, OA system, training platform, project management system, and internal IM system to collect basic information, performance data, training data, and behavioral data of the evaluation subjects; External data channels: Connect with third-party professional credit platforms, industry skills certification agencies, and professional databases to collect data such as the professional credit records, industry skills certifications, external project experience, and industry reputation of the evaluated individuals; S2.2 standardizes the collected multi-source heterogeneous data, including: Data cleaning: Remove invalid data, use interpolation to supplement missing key data, and normalize and correct data with spelling errors and inconsistent formats; Data classification and coding: Classify data according to the set of evaluation dimensions, encode unstructured data into text, standardize and quantify structured data, and unify data format and units; Data anonymization: Sensitive information of the evaluation object is anonymized, and sensitive fields are hidden using encryption algorithms to ensure data security and compliance with Level 3 Information Security Protection Certification and related data security standards.

[0007] Step S3: Dynamic weight allocation and multi-dimensional comprehensive scoring S3.1 Based on the evaluation scenario and job profile, it uses a combination of the analytic hierarchy process (AHP) and reinforcement learning to dynamically allocate the weights of each evaluation dimension and indicator: First, the initial weights of each evaluation dimension are determined by using the analytic hierarchy process (AHP) based on the priority of each dimension in the job competency profile. Then, through reinforcement learning algorithms, combined with historical evaluation data, business demand change data and evaluation result feedback data, the initial weights are dynamically adjusted so that the weight allocation matches the evaluation scenario and job requirements to a preset threshold, thus avoiding weight solidification. S3.2 Based on the standardized data, scores are assigned to each evaluation indicator. The scoring methods include: Quantitative indicator scoring: The interval mapping method is used to map the standardized quantitative data to a preset scoring interval, and the specific score is determined according to the data size; Qualitative indicator scoring: Natural language processing (NLP) technology is used to perform sentiment analysis and keyword extraction on unstructured text data. Combined with preset qualitative evaluation criteria, qualitative indicator scores are automatically generated. At the same time, an entry point for manual review is retained, allowing manual adjustment of scores. S3.3 Based on the weights of each dimension and indicator, the weighted summation method is used to calculate the comprehensive score of the evaluation object. At the same time, the sub-scores of each evaluation dimension are calculated to form a multi-dimensional scoring result, which clearly presents the strengths and weaknesses of the evaluation object.

[0008] Step S4: Evaluate the self-iterative optimization of the evaluation model S4.1 Construct an iterative database for the evaluation model to store historical evaluation data, comprehensive scoring results, and business application feedback; S4.2 Periodically iteratively optimize the evaluation model, specifically including: Calculate the deviation between the evaluation results and the feedback from business applications. If the deviation exceeds the preset threshold, analyze the reasons for the deviation and adjust the evaluation indicators, weight allocation rules, or scoring standards. Based on newly collected talent data, machine learning algorithms are used to train the evaluation model, update the model parameters, and improve the accuracy of the evaluation model. In light of industry development trends and organizational business adjustments, add or delete evaluation indicators, optimize the set of evaluation dimensions, and ensure the adaptability of the evaluation model. After S4.3 iteration optimization, a model iteration report is generated, recording the optimization content, optimization reasons and optimization effects, for users to view and review, ensuring that the iteration process is traceable.

[0009] Step S5: Generation and Output of Personalized Evaluation Report S5.1 Select a preset report template based on the evaluation scenario, job type, and user needs. The report template includes fixed modules and customizable modules. S5.2 analyzes data based on multi-dimensional scoring results, job competency profiles, and evaluation models, fills in the report template, and generates a personalized evaluation report, specifically including: Basic Information Module: Displays basic information about the evaluated candidate, their job title, and the evaluation scenario. Comprehensive scoring module: Displays the comprehensive score, score level, and comparison with the average level of the same position; Dimensional Analysis Module: Displays the sub-scores and indicator scores for each evaluation dimension, and uses visual charts to intuitively present the performance of the evaluated object in each dimension; Strengths and Weaknesses Module: Based on the scoring results, it automatically identifies the core strengths and weaknesses of the evaluated object and provides targeted analysis in combination with job requirements; Improvement and Potential Module: Generates personalized improvement suggestions based on the weaknesses of the evaluated object; and provides talent growth potential predictions and development path suggestions based on potential dimension scoring and machine learning predictions. S5.3 offers multiple report output formats, allowing users to edit, annotate, and export report content. It also supports batch generation of evaluation reports for multiple evaluation objects, improving report generation efficiency.

[0010] This invention also provides a talent evaluation report generation system for implementing the above-mentioned talent evaluation report generation method, comprising: The initialization module is used to receive evaluation scenario parameters input by the user, obtain the job type corresponding to the evaluation object, generate a job competency profile, and initialize the evaluation dimension set. The initialization module includes a scenario parameter receiving unit, a job profile generation unit, and an evaluation dimension initialization unit. The job profile generation unit has a built-in preset job classification library and can customize job types and competency requirements according to the organization's business needs.

[0011] The data acquisition and processing module includes a multi-source data acquisition unit and a data standardization processing unit: Multi-source data acquisition unit: Constructs internal and external data channels, connects to various data sources, realizes automatic acquisition of multi-source heterogeneous data, and supports two modes: timed acquisition and manual triggering acquisition. Data standardization processing unit: Used to clean, classify, encode, and de-identify the collected data, and output standardized evaluation data to ensure the accuracy, consistency and security of the data.

[0012] The comprehensive scoring module includes a weight allocation unit and a score calculation unit: Weighting Unit: Using a combination of analytic hierarchy process and reinforcement learning, the weights of each evaluation dimension and indicator are dynamically allocated based on the evaluation scenario and job profile, and manual weight adjustment is supported. Scoring Calculation Unit: Performs quantitative and qualitative scoring on standardized data, calculates comprehensive and sub-item scores based on weight allocation results, and generates multi-dimensional scoring results.

[0013] The model iteration module is used to build an evaluation model iteration database, store historical evaluation data and business application feedback, periodically iterate and optimize the evaluation model, update evaluation indicators, weight rules and model parameters, and generate model iteration reports to ensure the accuracy and adaptability of the evaluation model. The model iteration module includes a data storage unit, a deviation analysis unit, a model training unit and an iteration report generation unit.

[0014] The report generation and output module includes a template selection unit, a report filling unit, and an output unit: Template selection unit: Includes various report templates for different scenarios and job types, and supports user-defined templates; Report population unit: Based on multi-dimensional scoring results and analysis data, automatically populate the report template and generate a personalized evaluation report; Output unit: Supports multiple report formats, provides report editing, annotation and batch generation functions to meet diverse user needs.

[0015] The data security module is used to ensure the data security of the entire system, including a data encryption unit, an access control unit, and a logging unit. Data encryption unit: Encrypts and stores sensitive data during transmission to prevent data leakage; Access control unit: Set different user roles and assign different operation permissions to ensure data access security; Log recording unit: Records all system operations to form an operation log, which facilitates traceability and auditing.

[0016] The advantages of this invention compared to the prior art are: This invention innovatively achieves a deep integration of evaluation scenarios and job profiles. By dynamically generating job competency profiles, it enables evaluation dimensions and indicators to be precisely matched with job requirements, breaking the limitations of traditional evaluation standards that are fixed and singular, improving the relevance of evaluations, and solving the problem of low matching degree between existing evaluations and job requirements. This invention constructs a multi-source heterogeneous data acquisition and standardized processing mechanism, integrates internal multi-system data and external authorized data, solves the data fragmentation problem, improves the completeness of talent profiles, and ensures data security and compliance through data anonymization processing. This invention employs a dynamic weight allocation method that combines the Analytic Hierarchy Process (AHP) with reinforcement learning to achieve automatic adjustment and optimization of weights, avoiding subjective biases caused by manual weight allocation. It also incorporates dynamic iteration based on evaluation feedback to make the weight allocation more aligned with business needs. This invention introduces a self-iterative mechanism for the evaluation model, which periodically optimizes the model based on historical data and business feedback, ensuring that the evaluation system can adapt to organizational development and industry changes. This solves the problem that existing evaluation models cannot dynamically adapt and enables continuous optimization of the evaluation model. This invention supports personalized and batch report generation, provides a variety of report templates and output formats to meet the diverse needs of different evaluation scenarios and users, simplifies the evaluation process, reduces manual intervention, shortens the evaluation process time, and improves report generation efficiency. The system architecture of this invention is comprehensive, with data security protection mechanisms and operation traceability functions. It is applicable to talent management scenarios of various organizations, highly practical, and can be widely promoted and applied. It can effectively help organizations achieve accurate identification, scientific training and rational allocation of talents, and improve talent management efficiency. Attached Figure Description

[0017] Figure 1 This is a flowchart of a talent evaluation report generation method according to the present invention.

[0018] Figure 2 This is a system block diagram of a talent evaluation report generation system according to the present invention. Detailed Implementation

[0019] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present invention.

[0020] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the invention or its application or use.

[0021] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0022] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0023] The following detailed description of a talent evaluation report generation method and system according to the present invention, with reference to the accompanying drawings, provides further insight.

[0024] Combined with appendix Figure 1-2 The specific implementation process of the talent evaluation report generation method and system of the present invention is as follows: A method for generating talent evaluation reports includes the following steps: Step S1: Initialize the evaluation scenario and job profile S1.1 Receive evaluation scenario parameters input by the user. The evaluation scenario parameters include the purpose of the evaluation, such as recruitment, promotion, training, and inventory; the evaluation cycle; and the scope of the evaluation object, such as individual, team, or department. S1.2 Based on the preset job classification library, obtain the job type corresponding to the evaluation object, call the job profile generation module, and generate a job competency profile by combining the organization's business needs and the core responsibilities of the job. The job competency profile includes core competency dimensions, auxiliary competency dimensions, competency requirement thresholds, and priority of each dimension. The core competency dimensions are dynamically determined according to the job type. For example, the core dimensions of a technical position are professional skills and innovation ability, and the core dimensions of a management position are coordination ability and communication ability. S1.3 Based on the evaluation scenario parameters and job competency profile, initialize the evaluation dimension set. The evaluation dimension set includes basic information dimension, competency dimension, performance dimension, potential dimension, and behavior dimension. Each dimension contains several specific evaluation indicators, and the selection of indicators is precisely matched with the job profile.

[0025] Step S2: Multi-source heterogeneous data acquisition and standardization processing S2.1 Construct a multi-source data acquisition channel, which includes an internal data channel and an external data channel: Internal data channels: Connect to the organization's internal HR system, OA system, training platform, project management system, and internal IM system to collect basic information of the evaluated individuals: education, length of service, position, and job level; performance data: KPI completion status, project results, and reward and punishment records; training data: course learning duration, assessment results, and skills certifications; and behavioral data: number of cross-departmental collaborations, task response speed, and work output quality. External data channels: Connect with third-party professional credit platforms, industry skills certification agencies, and professional databases to collect data such as the professional credit records, industry skills certifications, external project experience, and industry reputation of the evaluated individuals; S2.2 standardizes the collected multi-source heterogeneous data, including: Data cleaning involves removing invalid data, such as missing values, outliers, and duplicate data; using interpolation to supplement missing key data; and normalizing and correcting data with spelling errors and inconsistent formats. For missing data interpolation, linear interpolation is used, with the following formula: ; in, The interpolation results for missing data. This is the previous valid data point before the missing data. This is the first valid data point after the missing data. This indicates the index location of the missing data. Data Classification and Coding: Data is classified according to the set of evaluation dimensions. Unstructured data, such as project summaries, evaluation opinions, and interview records, are text-coded. Structured data, such as performance scores and learning duration, are standardized and quantified to unify data format and units. Structured data standardization uses the min-max normalization formula to map the data to the interval [0, 1], as shown in the formula below: ;in, For standardized data, This is the original data. This represents the minimum value of all raw data under this indicator. This represents the maximum value of all raw data under this indicator; Data anonymization: Sensitive information of the evaluation subjects, such as ID numbers, contact information, and salaries, is anonymized. Encryption algorithms are used to hide sensitive fields, ensuring data security and compliance with Level 3 Information Security Protection Certification and related data security standards. Sensitive information encryption uses the AES encryption algorithm, with core encryption as follows: ;in, The encrypted ciphertext, For AES encryption function, Sensitive information in plaintext This is the encryption key.

[0026] Step S3: Dynamic weight allocation and multi-dimensional comprehensive scoring S3.1 Based on the evaluation scenario and job profile, it uses a combination of the analytic hierarchy process (AHP) and reinforcement learning to dynamically allocate the weights of each evaluation dimension and indicator: First, using the Analytic Hierarchy Process (AHP), the initial weights of each evaluation dimension are determined based on the priority of each dimension in the job competency profile. Within the AHP, a judgment matrix $$A$$ is constructed, and the weights $$w_i$$ of each dimension are calculated. The consistency check formula for the judgment matrix is ​​as follows: ; ; in, As a consistency indicator, To determine the largest eigenvalue of a matrix, To evaluate the number of dimensions, For consistency ratio, As the average random consistency index, according to The value is determined; when When the judgment matrix satisfies the consistency requirement, the weight allocation is effective; Then, through reinforcement learning algorithms, combined with historical evaluation data, data on changes in business needs, and feedback data on evaluation results, the initial weights are dynamically adjusted to ensure that the weight allocation matches the evaluation scenario and job requirements to a preset threshold, such as above 90%, thus preventing weight solidification. The reward function formula for weight adjustment in reinforcement learning is as follows: ; in, For weight The corresponding reward value, , Weighting coefficients ( ), To ensure the weighting matches the evaluation scenario and job requirements, The accuracy of the evaluation results corresponding to the weights is dynamically optimized by maximizing the reward value. S3.2 Based on the standardized data, scores are assigned to each evaluation indicator. The scoring methods include: Quantitative indicator scoring: An interval mapping method is used to map the standardized quantitative data to a preset scoring range (0-100 points), and the specific score is determined based on the data size; the mapping formula is as follows: ;in, To score quantitative indicators, This refers to standardized quantitative data; Qualitative Indicator Scoring: Natural Language Processing (NLP) technology is used to perform sentiment analysis and keyword extraction on unstructured text data. Combined with preset qualitative evaluation criteria, qualitative indicator scores are automatically generated. An entry point for manual review is also retained, allowing for manual adjustment of scores. The text sentiment score formula in sentiment analysis is as follows: ;in, To score the sentiment of the text, The number of keywords extracted, For the first The weight of each keyword, For the first The sentiment score of each keyword (positive is positive, negative is negative); the qualitative index score is obtained by mapping the sentiment score to the 0-100 range; S3.3 Based on the weights of each dimension and indicator, a weighted summation method is used to calculate the comprehensive score of the evaluated object. Simultaneously, the sub-scores for each evaluation dimension are calculated to form a multi-dimensional scoring result, clearly presenting the strengths and weaknesses of the evaluated object. The comprehensive scoring formula is as follows: ; in, For comprehensive scoring, To evaluate the number of dimensions, For the first The weights of each evaluation dimension, For the first The number of indicators under each evaluation dimension For the first The first dimension The weight of each indicator, For the first The first dimension The scoring for each indicator; the scoring formula for each dimension is as follows: ;in, For the first The evaluation dimensions are scored separately.

[0027] Step S4: Evaluate the self-iterative optimization of the evaluation model S4.1 Construct an iterative database for the evaluation model to store historical evaluation data, comprehensive scoring results, and business application feedback, such as the performance of talents after promotion and the job suitability of recruited talents. S4.2 Periodically iteratively optimize the evaluation model, specifically including: Calculate the deviation between the evaluation results and the feedback from business applications. If the deviation exceeds a preset threshold, such as 10%, analyze the reasons for the deviation and adjust the evaluation indicators, weighting rules, or scoring standards. The deviation is calculated using the mean squared error (MSE) formula: ;in, This is the deviation value. For the sample size, For the first The actual values ​​reported by the business application of each sample For the first Predicted values ​​of evaluation results for each sample; Based on the newly collected talent data, machine learning algorithms, such as XGBoost, are used to train the evaluation model, update the model parameters, and improve the accuracy of the evaluation model. The objective function formula of the XGBoost algorithm is as follows: ;in, Let be the objective function. This is the loss function (which measures the deviation between the predicted and actual values). For the number of decision trees, For the first The regularization term of the decision tree is used to prevent overfitting; In light of industry development trends and organizational business adjustments, add or delete evaluation indicators, optimize the set of evaluation dimensions, and ensure the adaptability of the evaluation model. After S4.3 iteration optimization, a model iteration report is generated, recording the optimization content, optimization reasons and optimization effects, for users to view and review, ensuring that the iteration process is traceable.

[0028] Step S5: Generation and Output of Personalized Evaluation Report S5.1 Select a preset report template based on the evaluation scenario, job type and user needs, such as recruitment version, promotion version and training version. The report template includes fixed modules (basic information, comprehensive score, dimensional analysis) and customizable modules (strengths analysis, weakness improvement suggestions, potential prediction). S5.2 analyzes data based on multi-dimensional scoring results, job competency profiles, and evaluation models, fills in the report template, and generates a personalized evaluation report, specifically including: Basic Information Module: Displays basic information about the evaluated candidate, their job title, and the evaluation scenario. The comprehensive scoring module displays the overall score, score level, and comparison with the average level for the same position; the comparison with the average level for the same position uses a deviation rate formula. in, The deviation rate between the overall score of the evaluated candidate and the average level of the same position. The average comprehensive score of all candidates in the same position; Dimensional Analysis Module: Displays the sub-scores and indicator scores for each evaluation dimension, and uses visual charts to intuitively present the performance of the evaluated object in each dimension; Strengths and Weaknesses Module: Based on the scoring results, it automatically identifies the core strengths and weaknesses of the evaluated object and provides targeted analysis in combination with job requirements; the standard for strength identification is an indicator score ≥ 85 points, and the standard for weakness identification is an indicator score ≤ 60 points; Improvement and Potential Module: Generates personalized improvement suggestions based on the weaknesses of the evaluated individuals; provides talent growth potential predictions and development path suggestions based on potential dimension scoring and machine learning predictions; potential prediction uses a linear regression formula: ; in, To score the potential for talent development, Scoring is done for each item in the potential dimension. , These are the regression coefficients, obtained through training with historical data; S5.3 offers multiple report output formats, such as PDF, Word, Excel, and online preview. It supports users in editing, annotating, and exporting report content, and also supports batch generation of evaluation reports for multiple evaluation objects, improving report generation efficiency.

[0029] This invention also provides a talent evaluation report generation system for implementing the above-mentioned talent evaluation report generation method, comprising: The initialization module is used to receive evaluation scenario parameters input by the user, obtain the job type corresponding to the evaluation object, generate a job competency profile, and initialize the evaluation dimension set. The initialization module includes a scenario parameter receiving unit, a job profile generation unit, and an evaluation dimension initialization unit. The job profile generation unit has a built-in preset job classification library and can customize job types and competency requirements according to the organization's business needs.

[0030] The data acquisition and processing module includes a multi-source data acquisition unit and a data standardization processing unit: Multi-source data acquisition unit: Constructs internal and external data channels, connects to various data sources, realizes automatic acquisition of multi-source heterogeneous data, and supports two modes: timed acquisition and manual triggering acquisition. Data standardization processing unit: This unit is used to clean, classify, encode, and de-identify the collected data, and output standardized evaluation data to ensure the accuracy, consistency, and security of the data. The unit has built-in linear interpolation formula, min-max normalization formula, and AES encryption formula to automatically complete the data standardization process.

[0031] The comprehensive scoring module includes a weight allocation unit and a score calculation unit: Weighting Unit: This unit uses a combination of analytic hierarchy process (AHP) and reinforcement learning to dynamically allocate the weights of each evaluation dimension and indicator based on the evaluation scenario and job profile. It also supports manual weight adjustment. This unit has built-in AHP consistency test formula and reinforcement learning reward function formula to achieve automatic weight calculation and optimization. Scoring Calculation Unit: Performs quantitative and qualitative scoring on standardized data, calculates comprehensive and sub-item scores based on weight allocation results, and generates multi-dimensional scoring results; This unit has built-in quantitative indicator mapping formulas, qualitative indicator sentiment score formulas, and weighted summation formulas for comprehensive and sub-item scores, and automatically completes the scoring calculation.

[0032] The model iteration module is used to build an evaluation model iteration database, store historical evaluation data and business application feedback, periodically iterate and optimize the evaluation model, update evaluation indicators, weight rules and model parameters, and generate model iteration reports to ensure the accuracy and adaptability of the evaluation model. The model iteration module includes a data storage unit, a deviation analysis unit, a model training unit and an iteration report generation unit. Among them, the deviation analysis unit has a built-in mean squared error (MSE) formula, and the model training unit has a built-in XGBoost algorithm objective function formula to realize iterative optimization of the model.

[0033] The report generation and output module includes a template selection unit, a report filling unit, and an output unit: Template selection unit: Includes various report templates for different scenarios and job types, and supports user-defined templates; Report Filling Unit: Based on multi-dimensional scoring results and analysis data, the unit automatically fills in the report template and generates a personalized evaluation report. This unit has built-in formulas for the deviation rate of the average level of the same position and talent growth potential prediction formulas to complete the filling of the core data in the report. Output unit: Supports multiple report formats, provides report editing, annotation and batch generation functions to meet diverse user needs.

[0034] The data security module is used to ensure the data security of the entire system, including a data encryption unit, an access control unit, and a logging unit. Data encryption unit: Encrypts and stores sensitive data to prevent data leakage; built-in AES encryption formula to encrypt sensitive information. Access control unit: Set different user roles, namely administrator, evaluator, and evaluated, and assign different operation permissions to ensure data access security; Log recording unit: Records all system operations, including data collection, scoring, report generation, and model iteration, forming an operation log for easy traceability and auditing.

[0035] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for generating talent evaluation reports, characterized in that: Includes the following steps: Step S1: Initialize the evaluation scenario and job profile. Receive the evaluation scenario parameters input by the user, obtain the job type corresponding to the evaluation object, generate a job competency profile, and initialize the evaluation dimension set based on the evaluation scenario parameters and job competency profile. Step S2: Multi-source heterogeneous data acquisition and standardization processing. Construct a multi-source data acquisition channel to collect multi-source heterogeneous data of the evaluation object. Clean, classify, encode and de-identify the collected data, and output standardized evaluation data. Step S3: Dynamic weight allocation and multi-dimensional comprehensive scoring. The method combines the analytic hierarchy process (AHP) and reinforcement learning to dynamically allocate the weights of each evaluation dimension and indicator based on the evaluation scenario and job profile. The evaluation indicators are scored based on standardized evaluation data, and the comprehensive score and sub-item scores are calculated by combining the weights. Step S4: Evaluate the self-iterative optimization of the evaluation model, build an evaluation model iteration database, and periodically adjust the evaluation indicators, weight allocation rules and model parameters based on historical evaluation data and business application feedback, iteratively optimize the evaluation model and generate a model iteration report; Step S5: Personalized evaluation report generation and output. Select a preset report template according to the evaluation scenario and job type. Fill the template based on the comprehensive score, sub-scores and job competency profile to generate a personalized evaluation report and support multi-format output.

2. The method for generating a talent evaluation report according to claim 1, characterized in that: In step S1, the evaluation scenario parameters include the evaluation purpose, evaluation cycle, and evaluation object scope; the job competency profile includes core competency dimensions, auxiliary competency dimensions, competency requirement thresholds, and priority of each dimension, wherein the core competency dimensions are dynamically determined according to the job type; the evaluation dimension set includes basic information dimensions, competency dimensions, performance dimensions, potential dimensions, and behavioral dimensions, and each dimension contains several specific evaluation indicators that match the job profile.

3. The method for generating a talent evaluation report according to claim 2, characterized in that: In step S2, the multi-source data acquisition channels include internal data channels and external data channels; the internal data channels connect to the organization's internal HR system, OA system, training platform, project management system and internal IM system to collect the basic information, performance data, training data and behavioral data of the evaluation subjects; the external data channels connect to third-party professional credit platforms, industry skills certification institutions and professional databases to collect their professional credit records, industry skills certification and external project experience data after authorization by the evaluation subjects.

4. The method for generating a talent evaluation report according to claim 3, characterized in that: In step S3, the dynamic allocation of weights specifically includes: determining initial weights based on the priority of each dimension in the job competency profile using the analytic hierarchy process; dynamically adjusting the initial weights using a reinforcement learning algorithm, combined with historical evaluation data, business requirement change data, and evaluation result feedback data, so that the weight allocation matches the evaluation scenario and job requirements to a preset threshold; the scoring method includes quantitative indicator scoring and qualitative indicator scoring, wherein the quantitative indicators are mapped to a preset scoring range using an interval mapping method, and the qualitative indicators are generated using natural language processing technology combined with preset evaluation standards, while retaining an entry point for manual review.

5. The method for generating a talent evaluation report according to claim 4, characterized in that: In step S4, the iterative optimization of the evaluation model specifically includes: calculating the deviation between the evaluation results and the feedback from business applications; if the deviation exceeds a preset threshold, analyzing the reasons for the deviation and adjusting the evaluation indicators, weight allocation rules, or scoring standards; training the evaluation model using machine learning algorithms based on newly collected talent data and updating the model parameters; optimizing the set of evaluation dimensions in conjunction with industry development trends and organizational business adjustments; and generating a model iteration report after iterative optimization, recording the optimization content, reasons, and effects.

6. The method for generating a talent evaluation report according to claim 5, characterized in that: In step S5, the report template includes a fixed module and a customizable module. The fixed module includes basic information, comprehensive score, and dimensional analysis. The customizable module includes strengths analysis, suggestions for improvement of weaknesses, and potential prediction. The output format of the evaluation report includes PDF, Word, Excel, and online preview, and supports report editing, annotation, and batch generation of reports for multiple evaluation objects.

7. A talent evaluation report generation system, used to implement the talent evaluation report generation method according to any one of claims 1-6, characterized in that, include: The initialization module is used to receive evaluation scenario parameters input by the user, obtain the job type corresponding to the evaluation object, generate a job competency profile, and initialize the evaluation dimension set. The data acquisition and processing module includes a multi-source data acquisition unit and a data standardization processing unit. The multi-source data acquisition unit is used to construct a multi-source data acquisition channel and acquire multi-source heterogeneous data. The data standardization processing unit is used to clean, classify, encode, and de-identify the acquired data. The comprehensive scoring module includes a weight allocation unit and a score calculation unit. The weight allocation unit is used to dynamically allocate the weights of each evaluation dimension and indicator using a combination of the analytic hierarchy process and reinforcement learning. The score calculation unit is used to score the standardized evaluation data and calculate the comprehensive score and the individual item scores. The model iteration module is used to build an evaluation model iteration database, periodically iterate and optimize the evaluation model, and generate model iteration reports. The report generation and output module includes a template selection unit, a report filling unit, and an output unit. The template selection unit provides preset report templates and supports custom templates. The report filling unit fills templates to generate personalized evaluation reports. The output unit supports multi-format report output, report editing, annotation, and batch generation. The data security module is used to encrypt, store, and transmit sensitive data, set different user roles and operation permissions, and record all operations to form a traceable operation log.

Images