An employee evaluation method, device, equipment and medium

Abstract

This invention relates to the field of enterprise human resource management, and discloses an employee assessment method, apparatus, equipment, and medium, comprising: collecting indicator data of quantitative indicator-type employees and behavioral data of behavioral performance-type employees in a team; calculating the indicator achievement rate as the performance score of the quantitative indicator-type employees based on the indicator data of the quantitative indicator-type employees; analyzing the behavioral performance-type employees' behavioral data using a large model based on behavioral assessment rules to obtain their performance scores; and calculating the assessment results of all employees in the team based on team assessment rules, using a large model, and the performance scores of the quantitative indicator-type employees and behavioral performance-type employees. This solves the problems of strong subjectivity, delayed feedback, and limited dimensions in existing employee assessment technologies.

Description

Technical Field

[0001] This invention relates to the field of enterprise human resource management, and in particular to an employee assessment method, device, equipment and medium. Background Technology

[0002] In the field of corporate human resource management, employee performance appraisal is a core element for optimizing resource allocation and motivating employee growth, and it has a significant impact on corporate operational efficiency and development quality. However, traditional employee performance appraisal methods and existing related data-driven prediction schemes still have significant shortcomings.

[0003] In terms of objectivity, traditional performance appraisals often rely on monthly scores from superiors or 360-degree environmental assessments. The former is easily influenced by subjective factors such as the evaluator's recent impressions and personal preferences, while the latter often leads to distorted results due to interpersonal relationships and cognitive biases. Neither approach can accurately reflect the true work performance of employees, highlighting the problem of strong subjectivity.

[0004] Regarding the timeliness of assessments, existing assessment methods are mostly conducted after the end of the month or quarter, which is a post-event review assessment. Even some data-driven predictive models lack a real-time data absorption mechanism and cannot dynamically adapt to business changes, making it difficult to identify employee performance shortcomings in a timely manner and resulting in serious feedback delays. This not only hinders timely improvement by individual employees but also indirectly affects the overall efficiency improvement of the team and the company.

[0005] In terms of data coverage dimensions, existing data-driven prediction solutions are mostly limited to a few quantifiable data such as attendance and task completion, and do not fully integrate key behavioral information such as project collaboration traces, communication quality, innovation contributions, and work feedback. The data dimensions are singular, making it difficult to comprehensively and three-dimensionally depict the value of employees' work, resulting in insufficient accuracy and comprehensiveness of the evaluation results.

[0006] Therefore, existing technologies suffer from problems such as strong subjectivity in employee performance evaluation, delayed feedback, and a single dimension. Summary of the Invention

[0007] This invention provides an employee assessment method, apparatus, computer equipment, and storage medium to solve the problems of strong subjectivity, delayed feedback, and single dimension in existing employee assessments.

[0008] Firstly, it provides an employee performance evaluation method, including: Collect quantitative indicator data from employees and behavioral data from employees exhibiting behavioral performance within the team; The performance score of employees with quantitative indicators is calculated based on the indicator achievement rate of the quantitative indicator data. Based on behavioral assessment rules, a large model is used to analyze the behavioral data of employees with strong behavioral performance to obtain their performance scores. Based on team performance evaluation rules, a large model is used to calculate the performance evaluation results of all team members based on the performance scores of employees with quantitative indicators and employees with behavioral performance.

[0009] Optional, the collection of indicator data for employees with quantitative indicators includes: Structured data from various systems is obtained through a multi-system interface mechanism, and core indicator data is extracted from the structured data. Collect quantitative indicators related to employees' work processes and extract supplementary indicator data from the unstructured data; By integrating core indicator data with supplementary indicator data, complete indicator data for quantitative indicator-based employees is obtained.

[0010] Optional, the collection of behavioral data from performance-based employees includes: Structured data from various systems is obtained through a multi-system interface mechanism, and quantitative behavioral data is extracted from the structured data. Collect unstructured data related to the work process of employees with behavioral performance, and extract non-quantitative behavioral data from the unstructured data; By integrating quantitative and non-quantitative behavioral data, complete behavioral data of employees with behavioral performance characteristics can be obtained.

[0011] Optional team evaluation rules include: Based on team attribute characteristics, match the corresponding basic rules and combine the basic rules into team assessment rules; When team attributes change, the basic rules are adjusted according to the current team attributes, and the team assessment rules are updated based on the adjusted basic rules.

[0012] Optionally, based on team performance evaluation rules, using a large model, the performance evaluation results for all team members are calculated based on the performance scores of employees with quantitative indicators and those with behavioral performance. These results include: To obtain the cross-job performance balance coefficient, the performance scores of employees with quantitative indicators and employees with behavioral performance are normalized, and the assessment results are calculated in combination with the team assessment rules. The cross-job performance balance coefficient is calculated based on the industry performance benchmark of the same type of job and the historical performance distribution within the company.

[0013] Optionally, team performance evaluation rules include performance level distribution rules, specifically including: The percentage of each performance level is dynamically determined based on the current number of team members. When the change in team members exceeds the preset percentage, the percentage of each level is recalculated based on the new team members and pushed to the large model for performance evaluation calculation.

[0014] Optionally, based on behavioral assessment rules, a large-scale model can be used to analyze the behavioral data of performance-oriented employees to obtain their performance scores, including: Semantic parsing is performed on behavioral data to obtain structured semantic parsing results of behavioral data. Feature extraction is performed on the structured parsing results of behavioral semantics to obtain a sequence of employee performance feature vectors; The employee performance feature vector sequence and behavioral assessment rules are input into the large model. The large model calculates the matching degree between the employee performance feature vector sequence and the behavioral assessment rules to obtain the performance score of the behavioral performance type employee.

[0015] Secondly, an employee assessment device is provided, comprising: The data collection module is used to collect metric data from employees with quantitative indicators and behavioral data from employees with behavioral performance in the team. The first scoring module is used to calculate the achievement rate of quantitative indicator employees based on their indicator data, which serves as their performance score. The second scoring module is used to analyze the behavioral data of employees with behavioral performance based on behavioral assessment rules and a large model to obtain the performance score of employees with behavioral performance. The comprehensive assessment module is used to calculate the assessment results of all employees in the team based on the team assessment rules and a large model, according to the performance scores of employees with quantitative indicators and employees with behavioral performance.

[0016] Thirdly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the aforementioned employee assessment method.

[0017] Fourthly, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the steps of the aforementioned employee assessment method.

[0018] The aforementioned employee assessment methods, devices, equipment, and storage media utilize a large-scale model to replace subjective human evaluation. Performance scores are quantified by the achievement rate of indicators for employees with quantitative indicators, while performance scores for employees exhibiting behavioral performance are generated by the large-scale model analyzing behavioral data using preset rules. The entire process is based on objective data and clear rules, avoiding interference from human bias. Relevant data from both types of employees is collected, and performance scores and final assessment results are dynamically generated through real-time calculation of indicator achievement rates and real-time analysis of behavioral data by the large-scale model, breaking the time limitations of traditional post-assessment. It covers both quantitative results and process behavior, integrating multi-dimensional information to avoid the biased assessment caused by a single data dimension. This effectively solves the problems of strong subjectivity, delayed feedback, and single-dimensionality in existing employee assessment technologies. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating an embodiment of the employee assessment method of the present invention; Figure 2 This is a schematic diagram of an employee assessment device according to an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0022] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0023] It should also be understood that the term “and / or” as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0024] As used in this specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," or "in response to determination." Similarly, the phrase "if determined" or "if matched to [described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once matched to [described condition or event]," or "in response to matched to [described condition or event]."

[0025] Furthermore, in the description of this invention and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0026] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of the invention include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0027] It should be understood that the sequence number of each step in the following embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0028] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0029] Please see Figure 1 As shown, Figure 1 A flowchart illustrating the employee performance evaluation method provided in this embodiment of the invention includes the following steps: S11: Collect quantitative indicator data from employees and behavioral data from employees with strong behavioral performance in the team; S12: Calculate the target achievement rate based on the target data of employees with quantitative indicators as the performance score of employees with quantitative indicators; S13: Based on behavioral assessment rules, use a large model to analyze the behavioral data of employees with strong behavioral performance to obtain their performance scores. S14: Based on the team assessment rules, using a large model, calculate the assessment results of all team members according to the performance scores of quantitative indicator employees and behavioral performance employees.

[0030] For ease of understanding, the following explains some key terms in this embodiment: Employee performance appraisal methods refer to a systematic process and set of techniques used to evaluate the work performance, abilities, and contributions of employees within an organization or team. These methods aim to provide objective and comprehensive evaluation results to support human resource management decisions, such as salary adjustments, promotions, training, and development.

[0031] Employees whose performance is primarily measured and evaluated through quantifiable metrics are those whose work performance is mainly determined by specific, quantifiable indicators. Examples include sales figures for sales personnel, production output for production personnel, and the number of problems solved by customer service personnel.

[0032] Behavioral performance-based employees are those whose work performance is primarily assessed through non-quantifiable or difficult-to-quantify aspects such as their behavior, attitude, collaboration, and problem-solving abilities during the work process. Examples include the innovation capabilities of R&D personnel, the team leadership of managers, and the service awareness of administrative staff.

[0033] Indicator data refers to numerical information used to measure the performance of employees with quantifiable metrics. This data is usually objective, statistically verifiable, and directly reflects an employee's achievement of specific tasks or goals.

[0034] Behavioral data refers to information that reflects the specific behaviors, interactions, and outputs exhibited by performance-oriented employees during the work process. This data may include various forms such as text, voice, and images, and requires further analysis and interpretation.

[0035] The target achievement rate refers to the ratio between the actual completion rate of a quantitative target set by an employee and the target value within a specific time period. This ratio is usually expressed as a percentage and is used to measure an employee's performance level on quantitative goals.

[0036] Performance rating is a score derived from a quantitative assessment of an employee's work performance. This score comprehensively reflects the employee's performance level across different dimensions and is an important component of the employee evaluation results.

[0037] Behavioral assessment rules refer to a set of pre-defined standards, logic, and weights used to guide large-scale model analysis of behavioral data of employees and to evaluate their performance. These rules define which behaviors are positive and which are negative, and the degree of influence of different behaviors on performance scores.

[0038] A large model refers to a deep learning model with a large number of parameters and powerful learning capabilities, capable of processing and understanding complex unstructured data and extracting meaningful patterns and information from it. In this embodiment, the large model is used to analyze behavioral data and calculate assessment results.

[0039] Team performance evaluation rules refer to a set of standards and logic used to guide the calculation of performance evaluation results for all team members. These rules comprehensively consider the performance scores of employees with quantitative indicators and those with behavioral performance, and may also include factors such as overall team goals and collaborative contributions.

[0040] Performance appraisal results refer to the comprehensive evaluation of an employee's work performance within a specific appraisal period, which is the final output of the employee appraisal method. These results are typically presented in the form of grades, scores, or written descriptions, and are used to guide subsequent human resource management decisions.

[0041] First, this method involves collecting metric data from employees with quantitative performance indicators (MPIs) and behavioral data from employees with behavioral performance indicators (VPIs). In one implementation, MPI data can be collected manually, for example, by employees or their supervisors periodically entering data such as sales figures and project completion counts into a spreadsheet. Behavioral data can be obtained through manual observation and recording, for example, by managers periodically writing employee work logs or behavioral observation reports. This collected data will serve as the basis for subsequent performance evaluations.

[0042] Secondly, the achievement rate of the quantifiable indicator-based employee is calculated as their performance score based on the indicator data. Specifically, a target value can be pre-set for each quantifiable indicator-based employee, and then the actual collected indicator data is compared with the target value to directly calculate the achievement rate. For example, if a salesperson's target sales amount is 1 million yuan, and the actual sales amount is 800,000 yuan, then their achievement rate is 80%. This achievement rate is then directly used as the employee's performance score.

[0043] Furthermore, based on behavioral assessment rules, a large-scale model is used to analyze the behavioral data of employees with specific behavioral performance traits to obtain their performance scores. In one implementation, a series of behavioral assessment rules can be preset; for example, "proactive communication" can be defined as a positive behavior, and "procrastination" as a negative behavior. Then, the collected behavioral data can be reviewed and judged manually, and employees' behavioral performance can be scored according to the preset rules, resulting in a comprehensive performance score for each employee.

[0044] Finally, based on the team's performance evaluation rules, a large model is used to calculate the performance evaluation results for all team members based on the performance scores of employees with quantitative indicators and those with behavioral performance. Specifically, a fixed set of team performance evaluation rules can be preset. For example, the performance scores of employees with quantitative indicators and those with behavioral performance can be weighted and averaged according to fixed weights to obtain the final evaluation result for each employee. This rule typically remains unchanged after being set and is not adjusted according to changes in the team's situation.

[0045] The following specific example will provide a more in-depth explanation of the above technical solution: Suppose that in a company called "Company A", there is a sales team and a research and development team. The employees in the sales team are mainly quantitative indicator-based employees, whose performance is primarily measured by quantitative indicators such as sales volume and customer conversion rate. The employees in the research and development team are mainly behavioral performance-based employees, whose performance is more reflected in behavioral performance such as innovation ability, teamwork, and problem-solving ability.

[0046] To address the issues of strong subjectivity, delayed feedback, and limited dimensions in existing employee performance evaluations, Company A adopted the method provided in this embodiment.

[0047] First, during the data collection phase, for sales team employees with quantitative performance indicators, the system automatically collects data such as sales revenue, number of customers, and contract signing rate from the company's Customer Relationship Management (CRM) and Enterprise Resource Planning (ERP) systems. For R&D team employees with behavioral performance indicators, the system collects behavioral data such as team members' speeches in project discussions, code commit records, document writing, and meeting minutes from project management tools, internal communication platforms, and code version control systems. This data collection aims to comprehensively cover employees' work output and process performance.

[0048] Secondly, during the performance evaluation phase, for sales team employees with quantitative performance indicators, the system compares the collected sales data with preset sales targets and automatically calculates each salesperson's target achievement rate. For example, if a salesperson's sales target is 1 million yuan and they actually achieve 800,000 yuan, their target achievement rate is 80%. This achievement rate is directly used as the salesperson's performance score. For R&D team employees with behavioral performance indicators, the system inputs collected behavioral data (such as project discussion texts and code commit logs) into a large model pre-configured with behavioral assessment rules. This large model performs in-depth analysis of the behavioral data according to the rules, identifying employees' specific performance in innovation, collaboration, problem-solving, etc., and generating a performance score for each R&D employee accordingly. For example, the large model might identify that an R&D employee proposed multiple innovative solutions in project discussions and actively assisted other team members in solving technical problems, thus giving them a high behavioral performance score.

[0049] Finally, in the performance evaluation calculation phase, Company A's system, based on pre-defined team evaluation rules, utilizes a larger model to comprehensively consider the quantitative performance scores of sales team employees and the behavioral performance scores of R&D team employees. This larger model weights and integrates different types of performance scores based on factors such as overall team goals and cross-functional contribution balance, ultimately calculating the evaluation results for all team employees. For example, the larger model may adjust the individual performance scores of sales and R&D personnel based on the achievement of overall team sales targets and the progress of R&D projects to reflect the differences in their contributions to the overall team, thereby obtaining a more comprehensive and fair evaluation result.

[0050] Through the above process, the method of this embodiment can effectively integrate two types of data: quantitative indicators and behavioral performance. By leveraging the powerful analytical capabilities of large models, it reduces the subjectivity of manual evaluation and provides multi-dimensional and more timely performance feedback.

[0051] The aforementioned employee assessment methods, devices, equipment, and storage media utilize a large-scale model to replace subjective human evaluation. Performance scores are quantified by the achievement rate of indicators for employees with quantitative indicators, while performance scores for employees exhibiting behavioral performance are generated by the large-scale model analyzing behavioral data using preset rules. The entire process is based on objective data and clear rules, avoiding interference from human bias. Relevant data from both types of employees is collected, and performance scores and final assessment results are dynamically generated through real-time calculation of indicator achievement rates and real-time analysis of behavioral data by the large-scale model, breaking the time limitations of traditional post-assessment. It covers both quantitative results and process behavior, integrating multi-dimensional information to avoid the biased assessment caused by a single data dimension. This effectively solves the problems of strong subjectivity, delayed feedback, and single-dimensionality in existing employee assessment technologies.

[0052] In practice, ensuring that the collected indicator data is comprehensive, accurate, and effectively reflects employees' actual performance is a technical problem that needs to be solved. A single data collection method may not cover all key performance information, leading to biased or inaccurate performance evaluations, which in turn affects the accuracy of subsequent performance scores.

[0053] In response, this application further proposes a method for collecting indicator data for quantitative indicator-type employees, including: obtaining structured data from various systems through a multi-system interface mechanism, and extracting core indicator data from the structured data; collecting unstructured data related to the work process of quantitative indicator-type employees, and extracting supplementary indicator data from the unstructured data; and integrating the core indicator data and supplementary indicator data to obtain complete indicator data for quantitative indicator-type employees.

[0054] Specifically, a multi-system interoperability mechanism refers to a technical means that enables data exchange and interoperability between different information systems. This mechanism can be implemented in various ways, such as defining a unified application programming interface (API) for data calls between systems, using message queue middleware for asynchronous data transmission, or using a data integration platform for batch data extraction, transformation, and loading. Its purpose is to aggregate data scattered across various business systems within an enterprise.

[0055] Structured data refers to data organized in a predefined format, typically stored in relational databases, spreadsheets, or XML files. Retrieving structured data from various systems can be achieved by executing database queries, calling system-provided Web Service interfaces, or parsing data files in specific formats. The goal is to obtain easily quantifiable direct performance data generated by employees in various business activities. Extracting core indicator data from structured data refers to filtering out key data points that are crucial for employee performance evaluation based on pre-defined business rules or performance evaluation models. This can be accomplished by configuring data filtering rules, applying data mining algorithms, or using business intelligence tools for data slicing and drill-down.

[0056] Unstructured data refers to data that does not have a predefined data model or fixed format, such as text files, emails, chat logs, and audio / video files. Collecting unstructured data related to employee work processes can be achieved by integrating with internal enterprise document management systems, email systems, instant messaging platforms, or by using content crawling technologies to obtain publicly available or semi-public work-related information. The aim is to capture employee work details, processes, and contextual information that are difficult to capture in structured data. Extracting supplementary indicator data from unstructured data refers to using advanced data analysis techniques to identify, analyze, and quantify information that can assist in employee performance evaluation. This can be achieved using natural language processing technologies for text semantic analysis, sentiment recognition, and topic extraction, or by using image recognition and speech recognition technologies to extract relevant features from multimedia data.

[0057] Integrating core and supplementary indicator data refers to merging core indicators extracted from structured data with supplementary indicators extracted from unstructured data to form a unified and comprehensive dataset. This can be achieved by building a data warehouse or data lake, or by using data fusion algorithms to correlate and normalize data from different sources and of different types. Ultimately, the result is complete indicator data for employees with quantifiable performance metrics. This means that after the above collection, extraction, and integration processes, a comprehensive dataset is formed that includes all relevant performance dimensions of the employee and fully reflects their work performance, providing a solid data foundation for subsequent performance evaluation.

[0058] The following is a concrete example. As a specific implementation method, the multi-system integration mechanism can be represented as a microservices-based API gateway. This gateway is responsible for data interaction with the enterprise's sales management system, project management system, customer service system, etc. Through this API gateway, structured data from each system can be obtained, such as employee sales revenue and order volume data from the sales management system, and project completion rate and task delay rate data from the project management system. From this structured data, sales revenue, order volume, and project completion rate can be directly extracted as core indicator data according to a pre-set performance evaluation model. Simultaneously, unstructured data related to employee work processes can be collected, such as employee chat logs and meeting minutes obtained by integrating with internal enterprise collaboration platforms (such as WeChat Work and DingTalk), and work reports and technical solutions submitted by employees obtained from the document management system. From this unstructured data, Natural Language Processing (NLP) technology can be used to perform semantic analysis on chat logs to assess team collaboration efficiency, and keyword extraction on work reports to identify innovative points or problem-solving capabilities, thereby extracting collaboration efficiency scores, innovation contribution, and other supplementary indicator data. Finally, these core metrics (such as sales revenue and project completion rate) are integrated with supplementary metrics (such as collaboration efficiency score and innovation contribution) and stored in a unified data warehouse to form complete metric data for quantitative employees. This dataset will contain detailed performance data of employees across multiple dimensions.

[0059] Through the aforementioned technical solution, this application can comprehensively and accurately obtain complete indicator data for employees with quantifiable performance indicators. Specifically, it acquires structured data and extracts core indicator data through a multi-system integration mechanism, combines it with unstructured data related to the employee's work process and extracts supplementary indicator data, and then effectively integrates them, effectively solving the problems of partiality and inaccuracy that may exist in traditional data collection methods. This provides a solid and multi-dimensional foundation for subsequent calculation of performance scores based on the indicator data of employees with quantifiable performance indicators, thereby improving the objectivity and fairness of employee assessment and enabling performance evaluation to more accurately reflect the actual contributions of employees.

[0060] Correspondingly, in practice, behavioral data may come from diverse sources and take various forms. Failure to collect this data comprehensively and systematically may limit the accuracy and comprehensiveness of behavioral assessments, affecting the effectiveness of large-scale model analysis. To address this, this application further proposes the collection of behavioral data from performance-oriented employees, including: acquiring structured data from various systems through a multi-system interface mechanism, and extracting quantitative behavioral data from the structured data; collecting unstructured data related to the work process of performance-oriented employees, and extracting non-quantitative behavioral data from the unstructured data; and integrating quantitative and non-quantitative behavioral data to obtain complete behavioral data for performance-oriented employees.

[0061] Among them, the multi-system interoperability mechanism refers to a technology or method that enables data exchange and interoperability between different information systems. Its role is to ensure the breadth and diversity of behavioral data sources, avoid data silos, and provide a comprehensive data foundation for subsequent analysis. This mechanism can employ API (Application Programming Interface) integration, defining a unified interface specification to allow different systems to call and transfer data to each other; or it can use message queue middleware to achieve asynchronous communication and data transmission between systems, improving the efficiency and reliability of data processing.

[0062] Structured data refers to data with a fixed format and clear definition, typically stored in relational databases, spreadsheets, etc., making it easy to query and analyze. Its purpose is to provide directly quantifiable behavioral information as the basis for behavioral assessment. Examples include clock-in / out records in employee attendance systems, task completion status in project management systems, and customer interaction counts in CRM systems. Quantifiable behavioral data refers to employee behavioral data extracted from structured data that can be directly measured numerically. Its purpose is to provide objective and quantifiable evidence for behavioral assessment, facilitating statistical analysis and pattern recognition in large models. Examples include the number of late arrivals and early departures extracted from attendance systems, task completion rates extracted from project management systems, and sales revenue or customer visit volume extracted from sales systems.

[0063] Unstructured data refers to data without a predefined data model or fixed format, such as text, images, audio, and video. Its purpose is to supplement structured data by providing richer and deeper insights into employee behavior details and contextual information that structured data cannot capture. Examples include employee work logs, email correspondence, meeting minutes, instant messaging records, project documents, and customer feedback texts. Non-quantitative behavioral data refers to data extracted from unstructured data that is difficult to measure directly with numerical values ​​but reflects employee behavioral characteristics. Its purpose is to provide semantic-level behavioral information for larger models, helping to understand the motivations, processes, and quality of employee behavior, thus compensating for the shortcomings of purely quantitative data. Examples include descriptions of project challenges extracted from work logs, team collaboration communication content extracted from emails, and service attitude evaluations extracted from customer feedback.

[0064] The purpose of integrating quantitative and non-quantitative behavioral data is to unify and fuse behavioral data from different sources and in different formats, forming a comprehensive and consistent dataset that provides high-quality input for large-scale model analysis. This can be achieved through techniques such as data cleaning, data transformation, and data standardization, unifying heterogeneous data into a pre-defined data model; or by building a unified data lake or data warehouse to store structured and unstructured data together and link them through metadata management. Ultimately, this yields complete behavioral data on employees, ensuring the comprehensiveness and accuracy of behavioral assessments, providing sufficient information for in-depth analysis of large-scale models, and thus deriving more objective and fair performance scores.

[0065] The following example illustrates this. Suppose a company needs to evaluate the performance of its R&D department employees based on behavioral characteristics. First, through a multi-system integration mechanism, the system can connect to the company's Jira project management system, Git code repository, WeChat Work communication platform, and internal knowledge base. From the Jira system, structured data such as the number of tasks completed and defect repair rate can be obtained. This data, after processing, can be used as quantifiable behavioral data. From the Git code repository, structured data such as the frequency of code submissions and participation in code reviews can be obtained, also as quantifiable behavioral data. Simultaneously, the system collects unstructured data such as employee chat logs in WeChat Work and documents and comments published in the internal knowledge base. Through natural language processing technology, non-quantifiable behavioral data such as the number of times employees assisted colleagues in solving problems and the degree of active participation in discussions can be identified from the WeChat Work chat logs; non-quantifiable behavioral data such as the quality of knowledge shared and the level of detail in answers to others' questions can be extracted from the knowledge base documents. Subsequently, these quantifiable and non-quantifiable behavioral data obtained from different systems and data types are uniformly cleaned, formatted, and stored to ultimately form a complete behavioral data profile of the R&D employee, which is then used for subsequent performance evaluation in a larger model.

[0066] In some embodiments, this application further proposes team assessment rules, including: matching corresponding basic rules based on team attribute characteristics, and combining the basic rules into team assessment rules; when team attribute characteristics change, adjusting the basic rules according to the current team attribute characteristics, and updating the team assessment rules based on the adjusted basic rules.

[0067] Team attribute characteristics refer to the inherent, identifiable features of a team that influence its operation and performance, such as team size (number of people), team type (e.g., R&D team, sales team, operations team), project complexity, team maturity, and business objectives. These characteristics can be defined and identified through pre-set system tags or categories, or automatically identified and extracted by algorithms through analysis of historical team data (e.g., project reports, personnel composition, business output). Basic rules refer to the smallest, independently effective logical unit constituting the team assessment rules, such as the performance level distribution ratio for different team sizes, the weight allocation for different project types, and the assessment cycle setting for different business objectives. These basic rules can be pre-defined as a series of standardized rule templates, each corresponding to one or more team attribute characteristics, or rules applicable to specific scenarios can be extracted through expert experience or historical data analysis. Matching the corresponding basic rules aims to ensure that the selected assessment rules are consistent with the team's actual situation, improving the relevance and effectiveness of the assessment. This can be achieved by establishing a mapping table between team attribute characteristics and basic rules, allowing the system to query and retrieve the corresponding basic rules based on the current team attribute characteristics, or by using a rule engine or decision tree model to dynamically filter and match the most suitable basic rules based on the input of team attribute characteristics. Combining basic rules into team assessment rules involves integrating multiple independent basic rules into a complete and executable team assessment logic. For example, selected basic rules can be chained or paralleled using logical operators (such as AND, OR, IF-THEN) to form a comprehensive rule set, or weights can be assigned to different basic rules, and the final team assessment rules can be formed through weighted summation or other aggregation algorithms. When team attribute characteristics change, the system will trigger a rule adjustment and update mechanism to ensure the dynamic adaptability of the assessment rules. This can be manifested in the system automatically detecting whether team attribute characteristics have changed periodically or when specific events (such as changes in team members or project phases) occur, or allowing administrators to manually update team attribute characteristics, thereby triggering the rule adjustment process. Adjusting basic rules based on current team attributes involves modifying or replacing existing basic rules to adapt to new team circumstances. For example, the system might re-execute matching logic based on new team attributes, selecting new basic rules or modifying parameters of existing ones. Alternatively, a pre-set rule adjustment strategy might automatically trigger the addition, subtraction, or replacement of specific parameters in certain basic rules when specific team attributes change. Updating team performance evaluation rules based on the adjusted basic rules involves applying the adjusted rules to the team performance evaluation system, ensuring the real-time nature and effectiveness of the evaluation rules. For example, the system might automatically recombine the adjusted basic rules to generate new team performance evaluation rules and make them effective immediately, or provide a preview and approval process before updating to ensure the accuracy and compliance of rule changes.

[0068] This application's solution dynamically manages team performance evaluation rules by introducing team attribute characteristics, enabling flexible adjustments based on the team's actual situation. Specifically, the system first identifies the team's attribute characteristics and then matches and combines these characteristics to generate initial team performance evaluation rules. These rules are subsequently used by the large model to calculate the performance evaluation results for all team employees. When the team's attribute characteristics change—for example, changes in team size, project type, or business objectives—the system promptly detects these changes and adjusts the basic rules accordingly, updating the team performance evaluation rules. This dynamic adjustment mechanism ensures that the team performance evaluation rules respond in real-time to actual changes in the team, thereby improving the accuracy, fairness, and effectiveness of the evaluation results. It avoids evaluation bias caused by rule lag, ensuring that the large model always uses the performance evaluation rules that best match the team's current state when calculating performance results.

[0069] The following is a concrete example to illustrate this. Suppose a software development team, in the early stages of a project, is identified as a "small development team" and in the "early stage of the project." Based on these characteristics, the system will match "small team performance level distribution rules" (e.g., A-level 10%, B-level 30%, C-level 50%, D-level 10%) and "early stage project emphasis on collaboration and innovation rules" (e.g., collaboration weighted 30%, innovation weighted 20%) as basic rules, and combine these basic rules into the team's initial team assessment rules. As the project progresses, the team size increases, and the project enters the "project sprint phase," the system will detect changes in the team's attribute characteristics. At this point, the system will adjust the basic rules based on the new team attributes. For example, the "Small Team Performance Level Distribution Rules" will be changed to "Medium Team Performance Level Distribution Rules" (e.g., A-level 15%, B-level 35%, C-level 40%, D-level 10%), and the "Early Project Stage Focuses on Collaboration and Innovation Rules" will be changed to "Project Sprint Focuses on Delivery and Quality Rules" (e.g., Delivery weight 40%, Quality weight 30%). Finally, the system will update the team assessment rules based on the adjusted basic rules to ensure that subsequent assessment calculations accurately reflect the team's performance priorities during the sprint phase.

[0070] Through the above technical solution, this application ensures that the assessment rules always remain consistent with the actual attributes and characteristics of the team, thereby significantly improving the accuracy and fairness of the assessment results. This avoids assessment bias caused by rigid rules, enabling employee performance evaluations to more accurately reflect their contributions and performance within the team, thus effectively motivating employees and promoting the overall improvement of team performance.

[0071] In some embodiments, this application further proposes a method based on team assessment rules, using a large model to calculate the assessment results of all employees in the team based on the performance scores of quantitative indicator employees and behavioral performance employees. This includes: obtaining a cross-position performance balance coefficient, normalizing the performance scores of quantitative indicator employees and behavioral performance employees, and then calculating the assessment results in conjunction with team assessment rules. The cross-position performance balance coefficient is calculated based on the industry performance benchmarks for similar positions and the historical performance distribution within the enterprise.

[0072] The purpose of obtaining a cross-job performance balance coefficient is to introduce a moderating factor to balance the differences in performance scores between employees in different job types. This coefficient can be a numerical value used to adjust or weight performance scores for different job types, ensuring that the contributions of employees in different positions are fairly measured in the final assessment. It can be obtained through expert experience, such as subjective evaluation and assignment by human resources experts based on factors like the importance and difficulty of different job responsibilities; or automatically generated through data analysis models, such as using machine learning algorithms to analyze historical performance data, identify the inherent patterns in the performance distribution of different jobs, and generate the balance coefficient accordingly. The purpose of normalizing the performance scores of employees with quantitative indicators and those with behavioral performance is to eliminate differences in the dimensions and numerical ranges of performance scores between different types of employees, making them comparable. Normalization maps the original scores to a uniform range, such as 0 to 1 or 0 to 100, thereby preventing certain scores from dominating subsequent calculations due to their large original numerical ranges. Normalization methods can include min-max normalization, which linearly scales the data to a specified range; or Z-score normalization, which converts the data into a distribution with a mean of 0 and a standard deviation of 1. The description of how to combine performance scores, after balancing and normalizing them, with pre-defined team assessment rules to arrive at the final employee assessment results describes how to combine them with pre-defined team assessment rules to arrive at the final employee assessment results. Team assessment rules can include multiple dimensions, such as the achievement of overall team goals, team collaboration performance, and the proportion of individual contributions within the team. The combination method can be through weighted summation, i.e., assigning weights to different factors and then summing them up; or through decision trees, rule engines, etc., determining the final assessment results based on pre-defined logical judgments and conditional branches. The cross-job performance equilibrium coefficient is calculated based on industry performance benchmarks for similar positions and the company's internal historical performance distribution, further defining the calculation basis of the cross-job performance equilibrium coefficient and ensuring its objectivity and rationality. "Industry performance benchmarks for similar positions" refers to the performance level or standard generally achieved by positions with similar responsibilities within a specific industry. This can be obtained through market research, industry reports, professional consulting, etc. "Internal historical performance distribution" refers to the statistical distribution of employee performance scores across different positions within the company over a past period, including the mean, median, standard deviation, and quantiles. This can be obtained through statistical analysis of the company's historical performance evaluation data. Calculating a balance coefficient based on these two types of data ensures that the coefficient considers both general industry standards and the company's actual situation and historical performance, making the coefficient setting more scientific and fair.

[0073] This application's solution aims to address the discrepancies in dimensionality and standards among performance scores of different employee types by introducing a cross-job performance balancing coefficient and normalizing performance ratings, thereby ensuring the fairness and accuracy of the assessment results. Specifically, before calculating the assessment results for all team employees, a cross-job performance balancing coefficient is first obtained. This coefficient is not arbitrarily set but is based on in-depth analysis of performance benchmarks for similar positions in external industries and historical performance distributions within the company. This allows the coefficient to objectively reflect the inherent differences and relative value of performance measurement across different positions. Subsequently, this balancing coefficient is used to normalize the performance ratings of employees with quantitative indicators and those with behavioral performance. This process unifies performance ratings that might otherwise fall into different dimensional and numerical ranges onto a comparable scale, eliminating biases caused by different evaluation systems. The normalized performance ratings more accurately and fairly reflect the actual performance level of employees. Finally, these balanced and normalized performance ratings, combined with pre-defined team assessment rules, are input into a large model for comprehensive calculation, resulting in the assessment results for all team employees. In this way, the solution ensures that the contributions of employees in different positions can be evaluated within a unified and fair framework when conducting overall team assessments. This avoids the distortion of assessments caused by differences in the original scores, enabling the large model to make more effective comprehensive judgments and improving the accuracy and credibility of the assessment results.

[0074] The following example illustrates this. Assume a company has sales positions (quantitative indicators) and research and development (behavioral performance) positions. When calculating performance evaluations, the first step is to obtain cross-position performance equilibrium coefficients. For example, the company can analyze the average sales revenue of sales positions and the project success rate of research and development positions within the industry as industry performance benchmarks. Simultaneously, it can statistically analyze the average performance scores of sales staff and research and development staff over the past three years to obtain the company's historical performance distribution. Based on this data, a pre-defined algorithm is used to calculate the performance equilibrium coefficients between sales and research and development positions; for example, the coefficient for sales is 1.1, and the coefficient for research and development is 0.9. Next, the sales revenue achievement rate (performance score) of sales staff and the performance scores obtained from behavioral data analysis of research and development staff are normalized. For example, a min-max normalization method can be used to map all performance scores to the range of 0 to 100. Specifically, for sales staff, their original performance scores are multiplied by an equilibrium coefficient of 1.1 before normalization; for research and development staff, their original performance scores are multiplied by an equilibrium coefficient of 0.9 before normalization. Finally, these performance scores, adjusted and normalized by balancing coefficients, are input into the large model along with team assessment rules (e.g., overall team goal achievement rate accounts for 30%, individual performance accounts for 70%). The large model then comprehensively considers individual performance, team rules, and balanced job differences based on these inputs to calculate each employee's assessment result, such as A, B, C, or D grades.

[0075] Through the above technical solution, this application effectively solves the problem of unfair comparison of performance scores among employees in different job types due to differences in evaluation standards and metrics. By introducing a cross-job performance equilibrium coefficient calculated based on industry performance benchmarks and the company's internal historical performance distribution, and normalizing the performance scores, the performance scores of employees in different positions are measured on a unified and fair scale. This significantly improves the objectivity and accuracy of the assessment results, avoids assessment bias caused by differences in job characteristics, ensures the fairness of the assessment process, and thus more accurately identifies high-performing employees and motivates their all-round development.

[0076] In some embodiments, this application further proposes team assessment rules including performance level distribution rules, specifically including: dynamically determining the proportion of each performance level based on the current number of team members, and when the change in the number of team members exceeds a preset proportion, recalculating the proportion of each level based on the new team members and pushing it to the large model for assessment result calculation.

[0077] The performance rating distribution rule refers to the rule in employee evaluation that divides team members' performance results into several preset rating levels (e.g., Excellent, Good, Satisfactory, Needs Improvement), and sets corresponding limits on the number or proportion of people in each rating level. This rule aims to ensure a reasonable distribution of performance evaluation results, avoiding excessive concentration in any one rating level, thereby maintaining the differentiation and motivational effect of the evaluation. This can be achieved through preset percentage ranges (e.g., Excellent not exceeding 15%, Good not exceeding 35%), or by setting a maximum number of people in each rating level. The current team size refers to the actual number of members in a specific team during the evaluation period. This data is the basis for dynamically determining the percentage of people in each performance rating level. The current team size can be obtained by interfaceing with the company's human resource management system to synchronize team member onboarding, offboarding, and job transfer information in real time; or by periodically summarizing and updating team member lists through project management systems or attendance systems to obtain accurate team size data. Dynamically determining the percentage of people in each performance rating level means flexibly adjusting the proportion or number of people corresponding to different performance rating levels based on changes in the current team size. This dynamic adjustment mechanism aims to make performance distribution rules more aligned with the team's actual situation, avoiding rigid assessments caused by changes in team size. Implementation methods can include: pre-setting multiple grade distribution templates for different team size ranges, with the system automatically matching the appropriate template based on the current team size; or using mathematical functions or algorithms based on team size to calculate and generate reasonable proportions for each grade in real time. For example, for smaller teams, the proportion of excellent grades can be appropriately relaxed, while for larger teams, stricter proportion control may be required. The preset proportion refers to a threshold used to determine whether changes in team size warrant recalculating performance grade proportions. When the change in team size (e.g., relative to the previous assessment period or initial number) exceeds this preset proportion, the system will trigger a recalculation of grade proportions. This preset proportion can be a percentage value (e.g., 10%, 20%), representing a percentage increase or decrease in team size compared to the original number; or it can be an absolute value (e.g., a change of more than 5 people), representing the absolute increase or decrease in team size. The setting of this proportion should comprehensively consider factors such as team size stability, assessment sensitivity, and management costs. Recalculating the percentage of each performance level refers to the system dynamically determining the percentage of employees in each performance level based on the latest team size when the team size changes by more than a preset percentage. This process ensures that the performance level distribution rules remain consistent with the latest team size. The recalculation method is similar to the dynamic determination of the percentage of employees in each performance level; it can involve reselecting a matching level distribution template or rerunning a mathematical function or algorithm based on the team size to generate a new set of performance level percentages suitable for the current team size.Pushing data to the main model refers to transmitting the latest calculated percentage of employees in each performance level to the main model responsible for calculating the assessment results via data interfaces or message queues. When calculating the final assessment results, the main model will refer to these dynamically updated level distribution rules to ensure that the final assessment results align with the team's actual situation and management requirements. Data push can be implemented through various methods such as real-time API calls, scheduled data synchronization, or event-driven mechanisms to ensure that the main model can promptly obtain and apply the latest rules.

[0078] The following example illustrates this. Suppose a research and development team sets performance rating distribution rules at the beginning of the year: Excellent (A) 10%, Good (B) 30%, Satisfactory (C) 50%, and Needs Improvement (D) 10%. The team initially consists of 20 people. During the assessment period, due to changes in project requirements, the team size increases to 25 people, representing a change of (25-20) / 20 = 25%. If the preset ratio is set to 20%, this change in team size exceeds the preset ratio. The system will dynamically recalculate the percentage of each performance rating based on the current team size of 25 people. For example, the system might adjust the distribution according to the preset dynamic adjustment algorithm: Excellent (A) 12% (3 people), Good (B) 32% (8 people), Satisfactory (C) 48% (12 people), and Needs Improvement (D) 8% (2 people). This updated performance rating distribution rule will then be pushed to the larger model. When the large model calculates the performance evaluation results for all employees in the team, it will use this new set of level distribution rules, which is more in line with the current team size, to make the final level classification and output of the employee's performance scores.

[0079] Through the above technical solution, this application effectively addresses the problems of rigid performance rating distribution rules and inability to adapt to dynamic changes in team size in traditional assessment methods. By dynamically determining the proportion of each performance rating based on the current team size and setting preset proportions to trigger a recalculation mechanism, the flexibility and timeliness of the performance rating distribution rules are ensured. This allows the assessment results to more accurately reflect the team's actual performance level, avoiding unfair assessments or incentive imbalances caused by changes in team size. Simultaneously, pushing the dynamically updated rating proportions to the large model allows the model to fully consider the team's latest situation when calculating assessment results, thereby improving the scientific rigor and rationality of the assessment results and enhancing the adaptability and fairness of the assessment system.

[0080] In some of the embodiments described above in this application, behavioral data of employees with specific behavioral characteristics are collected, and these data are analyzed using a large model based on behavioral assessment rules to obtain performance scores. However, behavioral data often has characteristics such as diversity, unstructured nature, and complexity. Directly inputting raw behavioral data into a large model for analysis may cause the large model to have difficulty accurately understanding the deeper meaning of the behavior, thereby affecting the objectivity and accuracy of performance scores.

[0081] In response, this application further proposes a method for obtaining performance scores for employees based on behavioral assessment rules and using a large model to analyze their behavioral data. This method includes: performing semantic parsing on the behavioral data to obtain a structured semantic parsing result; extracting features from the structured semantic parsing result to obtain a sequence of employee performance feature vectors; and inputting the employee performance feature vector sequence and behavioral assessment rules into a large model, using the large model to calculate the matching degree between the employee performance feature vector sequence and the behavioral assessment rules to obtain the performance score for the employee.

[0082] Semantic parsing of behavioral data aims to transform raw, potentially unstructured behavioral data into a structured form with clear semantics. For example, for text-based behavioral data, Natural Language Processing (NLP) techniques such as lexical analysis, syntactic analysis, named entity recognition, and sentiment analysis can be used to extract key behavioral subjects, actions, objects, timing, results, and the emotions or intentions implied by the behavior. Furthermore, predefined behavioral pattern recognition rules or ontology mapping can be used to parse non-textual behavioral data (such as system logs and operation records) into structured event sequences. Feature extraction from the semantically structured behavioral data parsing results aims to extract key information representing employee performance at the semantic level and transform it into a numerical representation that can be processed by large models. This can include statistically analyzing the frequency, duration, and intensity of specific behaviors, or calculating the pattern matching degree in behavioral sequences. For example, features can be extracted such as employee participation in collaborative communication, problem-solving efficiency, the number of innovative suggestions, and compliance with rules and regulations. These features can be encoded into multi-dimensional vectors, forming a sequence of employee performance feature vectors for subsequent quantitative analysis. The system inputs employee performance feature vector sequences and behavioral assessment rules into a large-scale model. The model calculates the matching degree and leverages its powerful analytical capabilities to objectively evaluate employee performance. This large-scale model can be a pre-trained deep learning model fine-tuned for the behavioral assessment task, capable of understanding the various standards and weights defined in the behavioral assessment rules. By comparing the employee performance feature vector sequences with the behavioral assessment rules, the large-scale model can identify the degree of conformity and deviation between employee behavior and assessment standards, and comprehensively consider the impact of various features on performance, thereby outputting a quantitative performance score.

[0083] The following is a concrete example to illustrate this. Suppose a company needs to assess the collaborative abilities of employees in its R&D team based on behavioral performance. First, the system collects behavioral data such as task assignment records in project management tools, code commit logs, discussion content in internal communication platforms, and meeting minutes. Next, semantic analysis is performed on this behavioral data. For example, for text data in the communication platform, natural language processing technology can be used to identify behavioral events such as "asking questions," "providing solutions," "assisting others," and "conflict resolution," along with their related entities and sentiment tendencies, forming structured semantic analysis results, such as "Employee A raised a technical question about module Y to employee B in project X and received positive assistance from employee C." Subsequently, employee performance characteristics are extracted from these structured results. For example, the number of times an employee "proactively provides solutions," the frequency of "cross-departmental collaboration," and the proportion of "effective communication" can be counted, and these indicators can be combined into an employee performance characteristic vector sequence. Finally, this characteristic vector sequence, along with preset collaborative ability assessment rules (for example, the rules might define the weight of "proactively providing solutions" as W1, the weight of "cross-departmental collaboration frequency" as W2, etc.), is input into a finely tuned large model. Based on these features and rules, the large model calculates the matching degree of employees in terms of collaboration ability and finally outputs a quantitative performance score, such as 85 points.

[0084] In one embodiment, this application also provides an employee assessment device. For example... Figure 2 As shown, it includes a data collection module 21, a first scoring module 22, a second scoring module 23, and a comprehensive assessment module 24. Detailed descriptions of each functional module are as follows: The data collection module 21 is used to collect the indicator data of quantitative indicator employees and the behavioral data of behavioral performance employees in the team. The first scoring module 22 is used to calculate the indicator achievement rate as the performance score of the quantitative indicator employee based on the indicator data of the quantitative indicator employee. The second scoring module 23 is used to analyze the behavioral data of behavioral performance employees based on behavioral assessment rules and a large model to obtain the performance score of the behavioral performance employees. The comprehensive assessment module 24 is used to calculate the assessment results of all employees in the team based on the team assessment rules and using a large model, according to the performance scores of employees with quantitative indicators and employees with behavioral performance.

[0085] This invention also provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the aforementioned employee assessment method; to avoid repetition, this will not be described further. Alternatively, the electronic device can implement the functions of each module in this embodiment of the employee assessment device; this will also not be described further.

[0086] This invention also provides a readable storage medium storing a program. When the program is executed by a processor, it implements the aforementioned employee assessment method. To avoid repetition, this will not be described further. Alternatively, when the program is executed by a processor, it implements the functions of each module in this embodiment of the employee assessment device, which will also not be described further.

[0087] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

[0088] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0089] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

Claims

1. An employee performance evaluation method, characterized in that, include: Collect quantitative indicator data from employees and behavioral data from employees exhibiting behavioral performance within the team; The achievement rate of the quantitative indicator employees is calculated based on their indicator data as their performance score. Based on the behavioral assessment rules, the behavioral data of the employees with the behavioral performance characteristics are analyzed using a large model to obtain their performance scores. Based on the team assessment rules, using a large model, the assessment results of all team members are calculated according to the performance scores of employees with quantitative indicators and employees with behavioral performance.

2. The employee assessment method according to claim 1, characterized in that, The collection of indicator data for the quantitative indicator-based employees includes: Structured data from each system is obtained through a multi-system interface mechanism, and core indicator data is extracted from the structured data. Collect and quantify unstructured data related to the work process of employees with quantitative indicators, and extract supplementary indicator data from the unstructured data; By integrating the core indicator data with the supplementary indicator data, complete indicator data for quantitative indicator-based employees is obtained.

3. The employee assessment method according to claim 1, characterized in that, The collection of behavioral data from the aforementioned performance-based employees includes: Structured data from each system is obtained through a multi-system interface mechanism, and quantitative behavioral data is extracted from the structured data. Collect unstructured data related to the work process of employees with behavioral performance characteristics, and extract non-quantitative behavioral data from the unstructured data; By integrating the quantitative and non-quantitative behavioral data, complete behavioral data of employees with behavioral performance characteristics can be obtained.

4. The employee assessment method according to claim 1, characterized in that, The team assessment rules include: Based on team attribute characteristics, corresponding basic rules are matched, and the basic rules are combined into team assessment rules. When team attributes change, the basic rules are adjusted according to the current team attributes, and the team assessment rules are updated based on the adjusted basic rules.

5. The employee assessment method according to claim 1, characterized in that, The team-based assessment rules, utilizing a large model, calculate the assessment results for all team members based on the performance scores of quantitative indicator-based employees and behavioral performance-based employees, including: The cross-job performance balance coefficient is obtained, and the performance scores of quantitative indicator-based employees and behavioral performance-based employees are normalized. The assessment results are then calculated in combination with the team assessment rules. The cross-job performance balance coefficient is calculated based on the industry performance benchmark of the same type of job and the historical performance distribution within the company.

6. The employee assessment method according to claim 1, characterized in that, The team assessment rules include performance level distribution rules, specifically including: The percentage of each performance level is dynamically determined based on the current number of team members. When the change in team members exceeds the preset percentage, the percentage of each level is recalculated based on the new team members and pushed to the large model for performance evaluation calculation.

7. The employee assessment method according to claim 3, characterized in that, The process of obtaining performance scores for employees based on behavioral performance evaluation rules and using a large model to analyze their behavioral data includes: The behavioral data is semantically parsed to obtain the behavioral semantic structured parsing results; Feature extraction is performed on the structured parsing results of the behavioral semantics to obtain a sequence of employee performance feature vectors; The employee performance feature vector sequence and behavioral assessment rules are input into the large model. The large model calculates the matching degree between the employee performance feature vector sequence and the behavioral assessment rules to obtain the performance score of the behavioral performance type employee.

8. An employee assessment device, characterized in that, include: The data collection module is used to collect metric data from employees with quantitative indicators and behavioral data from employees with behavioral performance in the team. The first scoring module is used to calculate the indicator achievement rate as the performance score of the quantitative indicator employee based on the indicator data of the quantitative indicator employee. The second scoring module is used to analyze the behavioral data of the behavioral performance employees based on the behavioral assessment rules and using a large model to obtain the performance scores of the behavioral performance employees. The comprehensive assessment module is used to calculate the assessment results of all employees in the team based on the team assessment rules and a large model, according to the performance scores of employees with quantitative indicators and employees with behavioral performance.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the employee assessment method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the employee assessment method as described in any one of claims 1 to 7.

Images