Personnel resume data processing method and system

Abstract

The embodiment of the invention discloses a personnel resume data processing method and system, and relates to the technical field of data processing, and the method comprises the steps: obtaining an organization structure file and an unstructured free list file of a current personnel; calling a document analysis agent to analyze the free list file so as to extract and output structured data containing a plurality of preset fields; calling a plurality of semantic analysis agents, and performing parallel semantic analysis on the structured data from multiple dimensions through a parallel processing mode; and calculating semantic analysis results of the plurality of semantic analysis agents through a dynamic quantization rule engine according to pre-configured resume data quantization calculation logic, and generating resume data quantization values of all dimensions of the current personnel and / or an overall resume data quantization value. According to the embodiment of the invention, the problem that in the prior art, the key information extraction precision is low due to the fact that an unstructured cadre free-free table document cannot be automatically analyzed and a specialized collaborative semantic analysis capability combined with an organizational structure is lacked can be solved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for processing personnel resume data. Background Technology

[0002] In organizational and personnel management, the analysis and evaluation of cadres' resumes is a crucial step in cadre selection and talent pipeline development, and its efficiency and accuracy directly affect the scientific nature of personnel decisions. Currently, cadre resume analysis suffers from significant technical limitations, failing to meet the needs of organizational and personnel departments for efficient and accurate processing of resume data. Specific problems include: First, existing methods have low automation levels and struggle to process unstructured raw documents. Current mainstream resume analysis methods still rely on manual review or semi-automated table entry: while some human resource information systems support importing data from structured Excel spreadsheets and performing simple scoring based on static rules (such as age and education), they generally lack the ability to automatically parse unstructured personnel appointment and dismissal forms (such as PDF and Word formats). This necessitates a significant amount of manpower to manually convert resume information from paper or electronic unstructured documents into structured data, which is not only inefficient (manual processing of a single resume typically takes more than 30 minutes) but also prone to data entry errors due to human intervention, affecting the accuracy of subsequent analysis results.

[0003] Secondly, existing technical solutions suffer from serious deficiencies in semantic understanding and information extraction. Even when unstructured documents are converted into text using technologies such as Optical Character Recognition (OCR), traditional Natural Language Processing (NLP) methods or fixed-template-based parsing techniques struggle to accurately understand and extract the complex semantic information contained in a cadre's resume. They cannot match the work unit name to organizational structure information (such as the hierarchical relationship between the group headquarters, second-level units, and third-level units), nor can they accurately identify key policy-oriented resume elements such as professional lines (such as "information management" and "financial management"), grassroots experience, and cross-unit appointments. This results in a single dimension of resume information extraction, failing to provide comprehensive data support for the assessment of cadres' composite abilities. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a method and system for processing personnel resume data to solve the problems of low accuracy in extracting key information caused by the inability of existing technologies to automatically parse unstructured cadre appointment and dismissal forms and the lack of professional collaborative semantic analysis capabilities that combine organizational structure.

[0005] In a first aspect, embodiments of the present invention provide a method for processing personnel resume data, including: Obtain the organizational structure file and appointment / removal form file of the current personnel, wherein the appointment / removal form file is an unstructured document; The document parsing agent is invoked to parse the appointment and dismissal table file, so as to extract and output structured data containing multiple preset fields; Multiple semantic analysis agents are invoked to perform semantic analysis on the structured data from multiple dimensions through parallel processing, including parsing and matching the work unit name field in the structured data using the organizational structure file; The dynamic quantization rule engine calculates the semantic analysis results of the multiple semantic analysis agents based on the pre-configured resume data quantization calculation logic, generating the resume data quantization values ​​of each dimension and / or the overall resume data quantization value of the current person.

[0006] Furthermore, the document parsing agent is invoked to parse the appointment and dismissal table file, including: By calling the dynamic parsing service based on a large language model through the application programming interface, the appointment and removal table file is semantically understood and contextually reasoned to identify and extract data from the multiple preset fields. The extracted data is organized into a JavaScript object representation format that conforms to a predefined pattern; The structured data in the JavaScript object representation format is split according to the physical entity and stored in different database entities.

[0007] Furthermore, multiple semantic analysis agents are invoked to perform semantic analysis on the structured data from multiple dimensions through parallel processing, including: Call the organizational structure parsing agent to match the work unit name field in the structured data to the organizational structure tree in the organizational structure file, and output the level and unique number; The professional line identification agent is invoked, and based on the job description text field in the structured data, combined with the predefined professional line dictionary and large language model, one or more professional tags and function types are output; The intelligent agent for judging diverse experiences is invoked to determine, according to the configuration rules, whether each work experience in the resume field of the structured data belongs to the special experience defined by the policy; Call the grassroots experience analysis agent, and combine the organizational structure to parse the level to which the agent outputs, and determine whether the experience in the resume field meets the preset grassroots experience conditions; The qualification and honor analysis agent is invoked to perform standardized mapping and level determination on the professional and technical position field, academic institution field, and award and punishment field in the structured data.

[0008] Furthermore, the method also includes: Configure the logic for quantifying the resume data of the personnel to be evaluated in each project using a human-computer interaction method; Based on the configured quantitative calculation logic of the current personnel's resume data, determine: the multiple semantic analysis agents to be invoked, and / or the classification or recognition granularity of the semantic parsing models within the multiple semantic analysis agents.

[0009] Furthermore, the logic for quantifying the resume data of personnel to be evaluated in each project is configured through human-computer interaction, including: Through a visual interface, configure the quantitative calculation logic for the resume data of the personnel to be evaluated in each project; The configuration of the quantitative calculation logic includes the indicators used to quantify personnel resume data and the logical rules for calculating the quantitative value of each indicator; the logical rules include: the type of factor corresponding to each indicator, the logic for calculating the quantitative value of the indicator based on the factor, and the parameters required to execute the logic.

[0010] Furthermore, configure the quantification calculation logic for the resume data of the personnel to be evaluated in each project, including; Create a project quantitative record table; For each item in the project quantification record table: configure the corresponding set of personnel to be evaluated, and create a personnel quantification record table; For each person to be evaluated in the personnel quantification record table: configure a corresponding set of indicators and create an indicator quantification record table; For each indicator in the indicator quantification record table: configure the corresponding set of measurement standards and create a measurement standard quantification record table; Configure corresponding logical rules for each metric in the metric quantification record table; The project quantitative record table, personnel quantitative record table, indicator quantitative record table, and measurement standard quantitative record table together constitute a hierarchical quantitative data structure, which is used to store the quantitative values ​​of each level generated by the dynamic quantitative rule engine.

[0011] Furthermore, the process of performing calculations through the dynamic quantization rule engine also includes: Based on the task identifier of the resume data quantification, poll the project quantification record table and the personnel quantification record table to monitor the status of the resume data quantification task. When a failed resume data quantification task is detected, update the status of the associated indicator quantification record table and the measurement standard quantification record table to "failed" and record the reason for the failure. For the quantification tasks of the resume data to be executed, the tasks are scheduled to be executed in order of indicator priority. The logical rules are converted into executable statements for calculation, and the status of the quantification record table at each level is updated accordingly. After the task of quantifying resume data is completed, the quantified values ​​are summarized and the status of personnel quantification records is updated.

[0012] Secondly, embodiments of the present invention provide a personnel resume data processing system, the system comprising: The file acquisition module is used to acquire the organizational structure file and appointment / removal form file of the current personnel, wherein the appointment / removal form file is an unstructured document; A document parsing intelligent agent is used to parse the appointment and dismissal table file, extract and output structured data containing multiple preset fields; Multiple semantic analysis agents are used to perform semantic analysis on the structured data from multiple dimensions through parallel processing, including parsing and matching the work unit name field in the structured data using the organizational structure file; A central coordinator is used to schedule the concurrent execution of the multiple semantic analysis agents; The dynamic quantization rule engine is used to calculate the outputs of the multiple semantic analysis agents based on the pre-configured resume data quantization calculation logic, and generate the resume data quantization values ​​of each dimension and / or the overall resume data quantization value of the current person.

[0013] Furthermore, the document parsing intelligent agent is specifically used for: By calling the dynamic parsing service based on a large language model through the application programming interface, the appointment and removal table file is semantically understood and contextually reasoned to identify and extract data from the multiple preset fields. The extracted data is organized into a JavaScript object representation format that conforms to a predefined pattern; The structured data in the JavaScript object representation format is split according to the physical entity and stored in different database entities.

[0014] Furthermore, the plurality of semantic analysis agents include: An organizational structure parsing agent is used to match the work unit name field in the structured data to the organizational structure tree in the organizational structure file, and output the level and unique number. A professional line recognition intelligent agent is used to output one or more professional tags and functional types based on the job description text field in the structured data, combined with a predefined professional line dictionary and a large language model. A diversified experience judgment agent is used to determine, according to configurable rules, whether each work experience in the resume field of the structured data belongs to the special experience defined by policy. The grassroots experience analysis agent is used to determine whether the experience in the resume field meets the preset grassroots experience conditions by combining the organizational structure with the level to which the agent outputs. The qualification and honor analysis agent is used to standardize and map the professional and technical position field, academic institution field, and award and punishment field in the structured data and determine their level.

[0015] The technical solution provided by this invention, through deep integration of multi-agent collaboration and a dynamic rule engine, achieves automation and intelligence in the entire process of personnel resume data quantification, and has the following advantages: By calling the document parsing agent, it automatically converts unstructured appointment and dismissal forms into structured data with preset fields, breaking through the efficiency bottleneck and error limitations of traditional manual parsing; relying on the parallel processing mode of multiple semantic analysis agents, it can not only deeply mine the semantic value of structured data from multiple dimensions, but also accurately parse and match the work unit name field in conjunction with the organizational structure document, greatly improving the professionalism and accuracy of key information extraction; through the standardized calculation of semantic analysis results by the dynamic quantification rule engine, it can efficiently generate quantitative values ​​of personnel resume data in various dimensions and overall, providing objective, systematic and accurate data support for personnel resume evaluation, and comprehensively optimizing the automation, intelligence and professionalism of personnel resume data processing. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art 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.

[0017] Figure 1 This is a flowchart of a personnel resume data processing method provided in Embodiment 1 of the present invention; Figure 2 This is a flowchart of a process for configuring the quantitative calculation logic of resume data for personnel to be evaluated in various projects, as provided in Embodiment 2 of the present invention. Figure 3 This is a flowchart of the status monitoring and scheduling process of a dynamic rule engine for quantifying various history data tasks, provided in Embodiment 2 of the present invention. Figure 4 This is a schematic diagram of a personnel resume data processing system provided in Embodiment 3 of the present invention. Detailed Implementation

[0018] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0019] It should be understood that the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0020] The technical solution of the present invention will be described in detail below through various embodiments.

[0021] Example 1 This embodiment provides a method for processing personnel resume data, which can be executed by a corresponding personnel resume data processing system. See also... Figure 1 The method specifically includes the following steps 101-104.

[0022] Step 101: Obtain the organizational structure file and appointment / removal form file of the current personnel, where the appointment / removal form file is an unstructured document.

[0023] Specifically, its front-end interface can receive two types of files uploaded by users (usually personnel from the organization's human resources department) for the current personnel. The "current personnel" refers to those whose resume data is currently being quantified among all personnel to be evaluated. Personnel to be evaluated include, but are not limited to, cadres of enterprises, institutions, or government agencies who currently require resume data processing.

[0024] First, the system receives the organizational structure file of the current personnel uploaded by the user. This file can be in structured Excel format, defining the complete tree structure of the organization (e.g., a state-owned enterprise group). The file includes the names of all levels of subordinate units (e.g., second-level units, third-level units, etc.) from the group headquarters down, their hierarchical affiliation in the organizational tree, and a unique identifier for each unit (e.g., "02-001" represents a second-level unit). The organizational structure file serves as a benchmark for subsequent semantic analysis by the intelligent agent, forming the basis for accurately resolving the hierarchical structure and relationships of personnel's employing units. Of course, the uploaded organizational structure file can also be in other structured formats; this embodiment of the invention does not limit this.

[0025] Secondly, the system receives user-uploaded appointment and dismissal forms for current personnel. These can be standard format documents defined by the organization, but are unstructured or semi-structured documents in PDF or Word format. Although the content may follow certain formatting conventions, the computer system cannot directly read the semantic information contained within (e.g., which paragraphs describe work experience, which unit names belong to third-level units). The document contains comprehensive information about the personnel, such as basic personal information (name, date of birth), detailed educational background (full-time, part-time), a work resume ordered chronologically (start and end dates, work units, and positions for each period), professional and technical titles, and awards and honors received.

[0026] Step 102: Call the document parsing agent to parse the appointment and dismissal table file to extract and output structured data containing multiple preset fields.

[0027] Specifically, a specialized document parsing agent is invoked to perform the parsing task of the appointment and removal form file. This agent does not rely on traditional template matching or OCR (Optical Character Recognition) layout reconstruction techniques, but instead utilizes the powerful semantic understanding and contextual reasoning capabilities of its integrated LLM (Large Language Model) to deeply interpret the uploaded cadre appointment and removal form (PDF / Word format). For example, the parsing process includes: By calling the dynamic parsing service based on a large language model through the application programming interface, the appointment and dismissal form file is semantically understood and contextually reasoned to identify and extract data from multiple preset fields. The extracted data is organized into a JavaScript object representation format that conforms to a predefined pattern; The structured data in the JavaScript object representation format is split according to the physical entity and stored in different database entities.

[0028] The above parsing process invokes a dynamic parsing service based on a large language model through a defined application programming interface (e.g., a RESTful API). This service leverages its powerful natural language processing capabilities to perform deep semantic understanding and contextual reasoning on unstructured personnel appointment and dismissal forms, accurately identifying, locating, and extracting semantic information units from multiple preset fields such as name, date of birth, educational background, and work experience. Subsequently, the extracted semantic information units are assembled into a standardized JavaScript object representation format according to a pre-designed, business-logic-compliant structural template. This JavaScript object constitutes the structured data skeleton of the personnel resume. This step achieves the transformation from unstructured text to structured data with clear hierarchical relationships. After the format conversion, the complete JavaScript object representation format structured data is logically decomposed according to the entity relationship model of the data persistence layer. For example, data from different dimensions such as basic information, educational experience, and work resume are persisted to different database entities (tables). This not only optimizes the data storage structure but also provides reliable data support for the efficient parallel access and analysis of subsequent semantic analysis agents.

[0029] Step 103: Invoke multiple semantic analysis agents and perform parallel semantic analysis on the structured data from multiple dimensions through parallel processing.

[0030] After extracting the structured data of cadre resumes, the system enters the deep semantic parsing stage. In this embodiment, a central coordinator can uniformly schedule and coordinate tasks between the document parsing agents and the semantic analysis agents to achieve the extraction of structured data and semantic analysis. The central coordinator, as the system's command center, distributes the structured data obtained in step 102 to various functionally independent agents according to a preset task flow, and manages their concurrent execution and result retrieval. This parallel processing architecture not only significantly improves analysis efficiency, but more importantly, through the professional division of labor and collaboration among agents, it effectively decouples complex semantic judgment tasks (such as organizational relationship mapping, professional domain classification, and policy condition matching) that were originally coupled together. This significantly enhances the maintainability and scalability of the system while ensuring parsing accuracy.

[0031] The multiple semantic analysis agents include an organizational structure parsing agent, a professional line identification agent, a diversified experience judgment agent, a grassroots experience analysis agent, and a qualification and honor analysis agent. Each agent can be an independently deployed semantic parsing microservice that receives structured data and returns semantic analysis results through a unified API interface. Furthermore, the organizational structure parsing agent uses the organizational structure file received in step 101 to parse and match the work unit name field in the structured data.

[0032] For example, the parallel semantic analysis process includes the following (1)-(5).

[0033] (1) Call the organizational structure parsing agent to match the work unit name field in the structured data obtained in step 102 to the organizational structure tree in the organizational structure file, and output the level and unique number.

[0034] In this process, the organizational structure parsing agent focuses on the work unit name field, calls the complete organizational structure tree (containing full-level organizational information from top to bottom) obtained from the organizational structure file uploaded by the user, and matches the work unit where the cadre is employed to the corresponding node in the structure tree through string matching and hierarchical tracing algorithms. Finally, it outputs the clear level to which the cadre's work unit belongs (such as standardized hierarchical expressions such as "group headquarters", "second-level unit", "third-level unit") and a unique identifier number, providing the core organizational hierarchy basis for subsequent validity determination of experience.

[0035] (2) Call the professional line recognition intelligent agent, based on the job description text field in the structured data obtained in step 102, combined with the predefined professional line dictionary and large language model, to output one or more professional tags and function types.

[0036] In this process, the first step is to use a predefined professional category dictionary (a standard classification system and keyword library covering various industries and fields) to perform preliminary matching and filtering of the job description text fields within their respective professional categories. This step is similar to a rapid keyword search, quickly identifying professional terms explicitly mentioned in the job description text fields that closely match the dictionary (for example, the presence of keywords such as "financial statements" and "budget preparation" in the text will initially point to the "financial management" category), efficiently narrowing the scope of analysis and providing candidate directions for subsequent in-depth analysis.

[0037] However, real-world job descriptions are often complex and use flexible terminology, resulting in numerous situations where simple keyword matching is insufficient. In such cases, the professional-specific identification agent can leverage the powerful semantic understanding capabilities of a finely tuned large language model to enhance the accuracy of identification within a specific professional domain. The large language model can understand the contextual implications of job responsibilities, handle ambiguities encountered in initial matching (e.g., "responsible for architecture design" might fall under "information management" rather than "construction engineering"), summarize abstract expressions (e.g., "responsible for overall business management" can be summarized as "comprehensive management"), and identify composite functions (e.g., "responsible for both technical team management and market expansion" can identify multiple tags such as "technical management" and "marketing").

[0038] Furthermore, the professional line identification agent integrates the preliminary matching of "keyword signals" with the deep "semantic understanding" of LLM, and outputs one or more accurate and practical professional labels (such as "information management", "financial management", "engineering construction" etc.), and clarifies its functional type (such as "professional and technical", "general management" and "business operation").

[0039] Preliminary matching is a fast filtering process based on explicit rules, responsible for capturing explicit features; while enhanced semantic understanding is a deep discrimination process based on model intelligence, responsible for interpreting implicit intentions and eliminating ambiguity. The two work together to ensure a balance between efficiency and accuracy in specialized line recognition.

[0040] (3) Call the diversified experience judgment agent and, according to the configuration rules, judge whether each work experience in the resume field of the structured data obtained in step 102 belongs to the special experience defined by the policy.

[0041] In this process, the intelligent agent for judging diverse experiences conducts analysis based on a pre-set configurable rule base. The rule base may include policy-defined criteria for judging special experiences (such as keyword lists like "serving in government departments," "working for aid units," and "participation in major special projects," as well as duration thresholds). The intelligent agent for judging diverse experiences breaks down the content of the work resume fields in the structured data segment by segment, and determines whether each work experience belongs to the policy-defined special experiences based on the judgment criteria, and marks the judgment basis and results.

[0042] (4) Call the grassroots experience analysis agent, combine the organizational structure to parse the level to which the agent outputs, and determine whether the experience in the resume field meets the preset grassroots experience conditions.

[0043] In this process, the results of the grassroots experience analysis agent and the organizational structure parsing agent are linked. First, the hierarchical information of the work unit output by the agent can be obtained. Then, combined with the work resume fields, the corresponding tenure period, job level and other content of each experience, the experience is compared with the preset grassroots experience conditions. The logical verification algorithm is used to determine whether the experience meets the recognition criteria for grassroots experience, and the compliance status of the experience is obtained. The key judgment nodes that reach the conclusion are recorded to ensure that the process is traceable.

[0044] (5) Call the qualification and honor analysis agent to perform standardized mapping and level determination on the professional and technical position field, academic institution field and award and punishment status field in the structured data obtained in step 102.

[0045] In this process, the qualification and honor analysis agent processes the three core fields of professional and technical positions, educational institutions, and awards and punishments in the structured data. Professional and technical positions are standardized and mapped according to industry-unified standards (such as classifying and labeling different expressions of "senior engineer" and "professional senior engineer"). Educational institutions are classified and judged according to preset institutional classification standards (such as "Double First-Class Universities" and "Ordinary Undergraduate Universities"). Honors are classified according to the award level (such as national, provincial, municipal, etc.) and honor type. Finally, standardized and comparable qualification and honor analysis results are output.

[0046] It should be noted that the multiple semantic analysis agents mentioned are not limited to the types listed in the examples above. In practical applications, dedicated agents with different semantic analysis capabilities can be flexibly integrated and scheduled according to specific business scenarios and data parsing needs. For example, to meet the needs of in-depth analysis of experience in specific professional fields, a technical analysis agent can be introduced to identify and classify personnel's project experience and professional depth in specific technology stacks (such as artificial intelligence, blockchain, etc.); similarly, to assess personnel's international perspective and cross-cultural cooperation ability, an international experience analysis agent can be deployed, specifically for analyzing their overseas study, work, or international cooperation projects. In this design approach, as long as the semantic analysis needs are defined, the overall parsing capability of the system can be enhanced by connecting the corresponding agents without changing the core processing pipeline and the computational logic of the dynamic rule engine, thus improving the system's adaptability and maintainability in the face of diverse and evolving evaluation needs.

[0047] Step 104: Using the dynamic quantization rule engine, the semantic analysis results of the multiple semantic analysis agents are calculated according to the pre-configured resume data quantization calculation logic to generate the resume data quantization values ​​of each dimension and / or the overall resume data quantization value of the current person.

[0048] In this step, a dynamic quantification rule engine is used to automatically calculate the quantified value of the current personnel's resume data. The engine's operation relies on pre-configured resume data quantification calculation logic, which defines how to map the semantic analysis results of multiple specialized intelligent agents into specific numerical data. Specifically, the resume data quantification calculation logic includes, but is not limited to: various indicators used to quantify personnel resume data; specific measurement standards set for each indicator; specific, executable logical rules set for each measurement standard; and a summary strategy definition for calculating the quantification results of each indicator according to different dimensions or for overall calculation. In a specific embodiment of this invention, the quantification processing of resume data is specifically manifested in the process of scoring cadre resume data in a cadre evaluation scenario.

[0049] In practical applications, the logic for quantifying the resume data of personnel to be evaluated in each project can be configured through human-computer interaction. The configuration process can be executed by a dynamic quantification rule engine. The configuration process is highly flexible: it can customize unique resume data quantification logic for individual personnel, or it can batch configure unified resume data quantification logic for a group of personnel with the same characteristics (e.g., all personnel in the same company or department, divided into the same project), which can significantly improve the efficiency and consistency of the system in large-scale applications.

[0050] It's important to note that each calculation premise defined in the resume data quantification calculation logic must find its corresponding structured data field as input in the semantic analysis results. For example, when the resume data quantification calculation logic includes the calculation logic of "if the age is between 35 and 40 years old, then its corresponding quantification value is 5", its effective execution premise is that the structured data output by the document parsing agent must contain the "age" field and its accurate value successfully identified and extracted from the original appointment and dismissal table file. If this premise field is missing or fails to be parsed successfully, this rule will not be correctly triggered and executed due to the lack of input data. This strong dependency and consistency verification mechanism established between the data layer and the logic layer technically ensures that the dynamic quantification rule engine processes complete and valid input information, thereby guaranteeing the reliability of subsequent logical judgments and numerical calculations, and ultimately generating accurate and reliable dimensional quantification values ​​and overall quantification values.

[0051] Accordingly, as a preferred implementation, based on the pre-configured resume data quantification calculation logic, step 103 determines: the multiple semantic analysis agents to be invoked; and / or, the classification or recognition granularity of the semantic parsing models within the multiple semantic analysis agents to be invoked. Specifically, at the agent scheduling level, by parsing the resume data quantification calculation logic, the preset set of data fields necessary to complete the quantification of the current personnel's resume data is identified. Based on this field requirement, the central coordinator can dynamically select and invoke semantic analysis agents that can provide the corresponding field analysis results. For example, if the calculation logic only requires "organizational level" and "professional line" information, only the organizational structure parsing agent and the professional line recognition agent are invoked, without activating other agents unrelated to the current task (such as the qualification and honor analysis agent), thereby achieving on-demand resource allocation and improved processing efficiency. At the agent internal parsing level, configuration instructions can also be sent to the invoked semantic analysis agents according to the detailed requirements of the resume data quantification calculation logic to dynamically adjust the classification or recognition granularity of their internal semantic parsing models. For example, based on evaluation needs, the professional line identification agent can be instructed to further refine the broad category of "information management" into more refined subcategories such as "infrastructure management" and "data architecture," thereby enabling the analysis results to better adapt to the accuracy requirements of downstream quantitative rules.

[0052] This optimization scheme achieves closed-loop linkage between the dynamic rule engine and multi-agent semantic analysis. The dynamic rule engine not only undertakes subsequent quantitative calculation tasks but also guides the parsing strategies of each specialized intelligent agent at the front end. Essentially, it constructs an agile response mechanism of "parsing as a service + rules as configuration," transforming the original unidirectional data processing pipeline into a dynamically adjustable, bidirectional interactive intelligent system. When an organization's evaluation policies or calculation rules change, there is no need to reconstruct the system; simply updating the configuration rules drives the entire parsing and calculation chain to automatically adapt. This achieves rapid and accurate response from business needs to technical execution, greatly improving the system's adaptability and overall intelligence level in the face of policy evolution.

[0053] Furthermore, after step 104, a resume data quantification report for the current personnel can be generated based on the quantified values ​​of each dimension of the personnel's resume data and / or the overall quantified values ​​of the resume data. This report can list each dimension and its corresponding quantified values ​​of the resume data, as well as the final overall quantified values ​​of the resume data. It can also reference the semantic analysis results of the multiple dedicated intelligent agents, specifically listing the key factual fragments from the original resume on which these quantified values ​​are based, thereby ensuring that the quantification results have sufficient factual basis and interpretability.

[0054] Example 2 This embodiment optimizes the step of "configuring the quantitative calculation logic of the resume data of the personnel to be evaluated in each project in a human-computer interaction manner" based on the above embodiment one. This step specifically includes: Through a visual interface, configure the quantitative calculation logic for the resume data of the personnel to be evaluated in each project; The configuration of the quantitative calculation logic includes the indicators used to quantify personnel resume data and the logical rules for calculating the quantitative value of each indicator. The logical rules include: the type of the factor corresponding to each indicator, the logic for calculating the quantitative value of the indicator based on the factor, and the parameters required to execute the logic.

[0055] When configuring the quantitative calculation logic for resume data, the first step is to select the indicators used to quantify the personnel resume data. These indicators constitute the basic dimensions of personnel resume data quantification. After determining the quantification dimensions, the corresponding factor type needs to be configured for each indicator. Currently, five factor types are supported: dictionary type, dictionary value type, functional type, agent type, and conventional type. Each type has an independent configuration entity. Functional factors support two implementation modes: configuring SQL query logic in rule-data collision scenarios, and configuring a logic bean based on in-memory computation in data-rule collision scenarios. Finally, the specific logic for calculating the quantitative value of the indicator based on the selected factor is configured, along with the parameters required for the execution of this specific logic. For example, for agent type factors, the parameters required for the corresponding specific logic include key information such as the agent service address, input parameter format, output data type, and result storage location.

[0056] As a preferred embodiment, see [link to previous document]. Figure 2 Configure the logic for quantifying the resume data of the personnel to be evaluated in each project, including the following steps 201-205.

[0057] Step 201: Create a project quantitative record table.

[0058] This step is the foundational step in configuring the entire quantitative calculation logic, used to coordinate quantitative data from a project perspective. This record table will create independent record entries for each project, with each entry serving as the overall carrier of quantitative information for that project. Subsequent quantitative data related to that project, such as personnel to be evaluated, indicators, and measurement standards, will be aggregated into the corresponding project quantitative record in a correlated manner, providing a top-level framework for subsequent hierarchical configuration.

[0059] Step 202: For each item in the project quantification record table: Configure the corresponding set of personnel to be evaluated and create a personnel quantification record table.

[0060] After creating the project quantification record table, for each project, it's necessary to define the specific personnel whose resume data needs to be quantified, i.e., configure the set of personnel to be evaluated. Specifically, this personnel set can be selected through a visual interface. Then, create a corresponding personnel quantification record for each person in this set; all personnel quantification records together constitute the project's personnel quantification record table. This step achieves hierarchical decentralization from project to personnel, allowing the quantification logic configuration to be precisely linked to the specific evaluation object, thereby ensuring that the quantification data of each person to be evaluated has a clear attribution.

[0061] Step 203: For each person to be evaluated in the personnel quantification record table: configure the corresponding indicator set and create an indicator quantification record table.

[0062] For each employee to be evaluated recorded in the personnel quantification record table, the core dimensions used to measure their resume data need to be determined based on the evaluation requirements, and a corresponding set of indicators (such as years of work experience, project experience, skill certificates, etc.) needs to be configured. In practice, the indicator set can be determined directly through a visual interface, or a predefined indicator set can be obtained from a locally pre-defined rule information table based on the resume analysis standard identifier received through this interface. Then, a corresponding quantification record is created for each indicator, and these records are integrated to form the personnel's indicator quantification record table. This step binds personnel to specific evaluation dimensions, allowing the quantification logic configuration to move beyond "who to evaluate" to "from which aspects to evaluate."

[0063] Step 204: For each indicator in the indicator quantification record table: configure the corresponding set of measurement standards and create a measurement standard quantification record table.

[0064] To achieve refined evaluation, each indicator in the quantitative record table needs to be further refined into specific measurement standards, i.e., a corresponding set of measurement standards needs to be configured. For example, if "project experience" is an indicator, its measurement standards may include "number of core project participations" and "duration of experience as a project leader," etc. In practice, the set of measurement standards can be determined directly through a visual interface, or the corresponding set of measurement standards can be obtained from a locally pre-built rule information table based on the indicator identifier. Based on these measurement standards, corresponding quantitative records are created to form the quantitative record table for the measurement standards of that indicator. This step transforms abstract indicators into concrete and operable measurement criteria, providing a refined carrier for the subsequent configuration of logical rules.

[0065] Step 205: Configure the corresponding logical rules for each metric in the metric quantification record table.

[0066] After determining the specific metrics, logical rules for calculating their quantification values ​​need to be configured for each metric. These rules include factor type selection (e.g., character type, function type), calculation logic (e.g., SQL query, in-memory computation Bean), and required parameters (e.g., agent service address, input / output format). These rules directly determine how to extract information from the resume data of the personnel being evaluated and calculate the quantification result of the metric, serving as the core execution basis for the quantification calculation logic.

[0067] Among them, the project quantitative record table, personnel quantitative record table, indicator quantitative record table and measurement standard quantitative record table together constitute a hierarchical quantitative data structure, which is used to store the quantitative values ​​of each level generated by the dynamic quantitative rule engine.

[0068] Of course, the same indicators, metrics, and logical rules can also be specified for different people working on the same project by default.

[0069] Accordingly, see Figure 3 The process of performing calculations through the dynamic quantization rule engine also includes: Step 301: Based on the quantified task identifier of the resume data, poll the project quantification record table and the personnel quantification record table to monitor the status of the resume data quantification task.

[0070] In this embodiment, a resume data quantification task refers to a task in which the dynamic quantification rule engine performs a complete resume data quantification calculation on a single individual. This task is marked by a unique resume data quantification task identifier and involves data records at multiple levels, including projects, personnel, indicators, and metrics. Based on the resume data quantification task identifier, the project quantification record table and the personnel quantification record table can be periodically polled to monitor the status of the task at each processing level (project level and personnel level) in real time, such as checking whether it is in the "pending execution," "in execution," "successful," or "failed" state.

[0071] Step 302: When a failed resume data quantification task is detected, update the status of the associated indicator quantification record table and the measurement standard quantification record table to "failed" and record the reason for the failure.

[0072] When monitoring detects a failure in a data quantification task, the system traces the error, updates the status of the parent records directly associated with the failure node (mainly the indicator quantification record table and the measurement standard quantification record table) to "failed", and records the specific reason for the failure in the log or related record fields.

[0073] Step 303: For the quantification task of the resume data to be executed, schedule the execution according to the priority of the indicators, convert the logical rules into executable statements for calculation, and update the status of the quantification record table at each level accordingly.

[0074] For tasks with a status of "pending execution," scheduling can be performed according to a preset priority order of metrics. Specifically, the configured logical rules are converted into statements that the system can directly execute (such as an SQL query or in-memory computation instruction). After the dynamic rule engine executes this statement, it updates the status and quantification values ​​of each level of quantification record table, from measurement standards to metrics, based on the results.

[0075] Step 304: After the resume data quantification task is completed, summarize the quantified values ​​and update the personnel quantification record status.

[0076] Once all indicators for a particular individual (corresponding to a resume data quantification task) have been calculated, the quantified values ​​are aggregated to generate the individual's overall quantified value. Finally, the status of the individual's quantification record table is updated to "success" and the final quantification result is written.

[0077] Example 3 This embodiment provides a personnel resume data processing apparatus, which can be used to execute the personnel resume data processing method described in this embodiment of the invention, and can be implemented by software and / or hardware. See also Figure 4 The device specifically includes the following units: The file acquisition module 401 is used to acquire the organizational structure file and appointment and dismissal form file of the current personnel, wherein the appointment and dismissal form file is an unstructured document; The document parsing intelligent agent 402 is used to parse the appointment and dismissal table file, extract and output structured data containing multiple preset fields; Multiple semantic analysis agents 403 are used to perform semantic analysis on the structured data from multiple dimensions through parallel processing, including parsing and matching the work unit name field in the structured data using the organizational structure file; Central coordinator 404 is used to schedule the concurrent execution of the multiple semantic analysis agents; The dynamic quantification rule engine 405 is used to calculate the outputs of the multiple semantic analysis agents according to the pre-configured resume data quantification calculation logic, and generate the resume data quantification values ​​of each dimension and / or the overall resume data quantification value of the current person.

[0078] For example, document parsing agent 402 is specifically used for: By calling the dynamic parsing service based on a large language model through the application programming interface, the appointment and removal table file is semantically understood and contextually reasoned to identify and extract data from the multiple preset fields. The extracted data is organized into a JavaScript object representation format that conforms to a predefined pattern; The structured data in the JavaScript object representation format is split according to the physical entity and stored in different database entities.

[0079] For example, multiple semantic analysis agents 403 include: An organizational structure parsing agent is used to match the work unit name field in the structured data to the organizational structure tree in the organizational structure file, and output the level and unique number. A professional line recognition intelligent agent is used to output one or more professional tags and functional types based on the job description text field in the structured data, combined with a predefined professional line dictionary and a large language model. A diversified experience judgment agent is used to determine, according to configurable rules, whether each work experience in the resume field of the structured data belongs to the special experience defined by policy. The grassroots experience analysis agent is used to determine whether the experience in the resume field meets the preset grassroots experience conditions by combining the organizational structure with the level to which the agent outputs. The qualification and honor analysis agent is used to standardize and map the professional and technical position field, academic institution field, and award and punishment field in the structured data and determine their level.

[0080] For example, the dynamic quantification rule engine 405 is also used to: configure the quantification calculation logic of the resume data of the personnel to be evaluated in each project in a human-computer interaction manner; The central coordinator 404 is also used to determine, based on the quantitative calculation logic of the current personnel's resume data configured by the dynamic quantitative rule engine 405, the classification or recognition granularity of the multiple semantic analysis agents to be invoked and / or the semantic parsing models within the multiple semantic analysis agents.

[0081] For example, the dynamic quantification rule engine 405 is used to configure the quantification calculation logic of the resume data of the personnel to be evaluated in each project in a human-computer interaction manner, specifically including: Through a visual interface, configure the quantitative calculation logic for the resume data of the personnel to be evaluated in each project; The configuration of the quantitative calculation logic includes: the indicators used in quantifying personnel resume data and the logical rules for calculating the quantitative value of each indicator; the logical rules include: the type of factor corresponding to each indicator, the logic for calculating the quantitative value of the indicator based on the factor, and the parameters required to execute the logic.

[0082] For example, the dynamic quantification rule engine 405 is used to configure the quantification calculation logic of the resume data of the personnel to be evaluated in each project, specifically including: Create a project quantitative record table; For each item in the project quantification record table: configure the corresponding set of personnel to be evaluated, and create a personnel quantification record table; For each person to be evaluated in the personnel quantification record table: configure a corresponding set of indicators and create an indicator quantification record table; For each indicator in the indicator quantification record table: configure the corresponding set of measurement standards and create a measurement standard quantification record table; Configure corresponding logical rules for each metric in the metric quantification record table; The project quantitative record table, personnel quantitative record table, indicator quantitative record table, and measurement standard quantitative record table together constitute a hierarchical quantitative data structure, which is used to store the quantitative values ​​of each level generated by the dynamic quantitative rule engine.

[0083] For example, the dynamic quantization rule engine 405's calculation process also includes: Based on the task identifier of the resume data quantification, poll the project quantification record table and the personnel quantification record table to monitor the status of the resume data quantification task. When a failed resume data quantification task is detected, update the status of the associated indicator quantification record table and the measurement standard quantification record table to "failed" and record the reason for the failure. For the quantification tasks of the resume data to be executed, the tasks are scheduled to be executed in order of indicator priority. The logical rules are converted into executable statements for calculation, and the status of the quantification record table at each level is updated accordingly. After the task of quantifying resume data is completed, the quantified values ​​are summarized and the status of personnel quantification records is updated.

[0084] The personnel resume data processing device in this embodiment can realize the complete set calculation method described in any of the foregoing embodiments. Its implementation principle and corresponding technical effects are basically the same, and will not be repeated here.

[0085] In summary, the technical solutions provided by the embodiments of the present invention have the following technical advantages: First, parsing accuracy and processing efficiency are improved simultaneously. By decoupling the complex task of quantifying cadre resume data into specialized intelligent agents such as organizational structure parsing, professional line identification, and policy element judgment, and with the help of a microservice collaborative architecture and a central coordination mechanism, the system accurately solves core pain points such as fuzzy matching of nested organizations and bias in the identification of key policy elements. This significantly reduces semantic ambiguity in document parsing, while the modular decomposition of tasks improves the convenience of subsequent system maintenance and iteration.

[0086] Secondly, the evaluation capability combines flexibility and accuracy. Relying on the closed-loop linkage mechanism between the dynamic rule engine and document parsing results, it not only supports the flexible definition and automated calculation of complex evaluation indicators such as "more than three years of management experience in third-level units", but also guides the intelligent agent to dynamically adjust semantic analysis strategies (such as optimizing the granularity of professional line classification) through preset rules, forming a virtuous cycle of "parsing output - rule evaluation - strategy iteration", ensuring the accuracy and adaptability of the evaluation results.

[0087] Third, the semantic understanding is deeply aligned with the needs of the scenario. An innovative three-in-one semantic modeling framework of "organizational level - professional line - policy indicators" is constructed, which is specifically adapted to the characteristics of complex organizational structure and high coupling of policy semantics in documents. It effectively avoids ambiguity problems in the parsing process, ensuring the accuracy of key resume information extraction and enriching the dimensional coverage of cadre evaluation.

[0088] Fourth, policy adaptability and system scalability are significantly enhanced. By adopting a design approach of "configuration-based adjustment" instead of "code-level modification," the system can quickly respond to the adjustment needs of organizational cadre selection policies without reconstructing the system architecture or retraining the intelligent agent model. It can flexibly adapt to the evaluation standards of large organizations of different periods and types, providing efficient, intelligent, logically interpretable, and agile iterative decision support for cadre selection, qualification review, and talent pipeline development.

[0089] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0090] In this embodiment of the invention, the term "and / or" describes the relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The character " / " generally indicates that the preceding and following associated objects have an "or" relationship.

[0091] The various embodiments in this specification are described in a related manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0092] In particular, the device embodiment is basically similar to the method embodiment, so the description is relatively simple. For relevant details, please refer to the description of the method embodiment.

[0093] For ease of description, the above apparatus is described by dividing it into various functional units / modules. Of course, in implementing this invention, the functions of each unit / module can be implemented in one or more software and / or hardware.

[0094] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0095] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for processing personnel resume data, characterized in that, include: Obtain the organizational structure file and appointment / removal form file of the current personnel, wherein the appointment / removal form file is an unstructured document; The document parsing agent is invoked to parse the appointment and dismissal table file, so as to extract and output structured data containing multiple preset fields; Multiple semantic analysis agents are invoked to perform semantic analysis on the structured data from multiple dimensions through parallel processing, including parsing and matching the work unit name field in the structured data using the organizational structure file; The dynamic quantization rule engine calculates the semantic analysis results of the multiple semantic analysis agents based on the pre-configured resume data quantization calculation logic, generating the resume data quantization values ​​of each dimension and / or the overall resume data quantization value of the current person.

2. The method as described in claim 1, characterized in that, The document parsing agent is invoked to parse the appointment and dismissal table file, including: By calling the dynamic parsing service based on a large language model through the application programming interface, the appointment and removal table file is semantically understood and contextually reasoned to identify and extract data from the multiple preset fields. The extracted data is organized into a JavaScript object representation format that conforms to a predefined pattern; The structured data in the JavaScript object representation format is split according to the physical entity and stored in different database entities.

3. The method as described in claim 1, characterized in that, Multiple semantic analysis agents are invoked to perform semantic analysis on the structured data from multiple dimensions through parallel processing, including: Call the organizational structure parsing agent to match the work unit name field in the structured data to the organizational structure tree in the organizational structure file, and output the level and unique number; The professional line identification agent is invoked, and based on the job description text field in the structured data, combined with the predefined professional line dictionary and large language model, one or more professional tags and function types are output; The intelligent agent for judging diverse experiences is invoked to determine, according to the configuration rules, whether each work experience in the resume field of the structured data belongs to the special experience defined by the policy; Call the grassroots experience analysis agent, and combine the organizational structure to parse the level to which the agent outputs, and determine whether the experience in the resume field meets the preset grassroots experience conditions; The qualification and honor analysis agent is invoked to perform standardized mapping and level determination on the professional and technical position field, academic institution field, and award and punishment field in the structured data.

4. The method as described in claim 1, characterized in that, The method further includes: Configure the logic for quantifying the resume data of the personnel to be evaluated in each project using a human-computer interaction method; Based on the configured quantitative calculation logic of the current personnel's resume data, determine: the multiple semantic analysis agents to be invoked, and / or the classification or recognition granularity of the semantic parsing models within the multiple semantic analysis agents.

5. The method as described in claim 4, characterized in that, Configure the quantitative calculation logic for the resume data of personnel to be evaluated in each project using a human-computer interaction method, including: Through a visual interface, configure the quantitative calculation logic for the resume data of the personnel to be evaluated in each project; The configuration of the quantitative calculation logic includes the indicators used to quantify personnel resume data and the logical rules for calculating the quantitative value of each indicator; the logical rules include: the type of factor corresponding to each indicator, the logic for calculating the quantitative value of the indicator based on the factor, and the parameters required to execute the logic.

6. The method as described in claim 5, characterized in that, Configure the quantification calculation logic for the resume data of the personnel to be evaluated in each project, including; Create a project quantitative record table; For each item in the project quantification record table: configure the corresponding set of personnel to be evaluated, and create a personnel quantification record table; For each person to be evaluated in the personnel quantification record table: configure a corresponding set of indicators and create an indicator quantification record table; For each indicator in the indicator quantification record table: configure the corresponding set of measurement standards and create a measurement standard quantification record table; Configure corresponding logical rules for each metric in the metric quantification record table; The project quantitative record table, personnel quantitative record table, indicator quantitative record table, and measurement standard quantitative record table together constitute a hierarchical quantitative data structure, which is used to store the quantitative values ​​of each level generated by the dynamic quantitative rule engine.

7. The method as described in claim 6, characterized in that, The process of performing calculations through the dynamic quantization rule engine also includes: Based on the task identifier of the resume data quantification, poll the project quantification record table and the personnel quantification record table to monitor the status of the resume data quantification task. When a failed resume data quantification task is detected, update the status of the associated indicator quantification record table and the measurement standard quantification record table to "failed" and record the reason for the failure. For the quantification tasks of the resume data to be executed, the tasks are scheduled to be executed in order of indicator priority. The logical rules are converted into executable statements for calculation, and the status of the quantification record table at each level is updated accordingly. After the task of quantifying resume data is completed, the quantified values ​​are summarized and the status of personnel quantification records is updated.

8. A personnel resume data processing system, characterized in that, include: The file acquisition module is used to acquire the organizational structure file and appointment / removal form file of the current personnel, wherein the appointment / removal form file is an unstructured document; A document parsing intelligent agent is used to parse the appointment and dismissal table file, extract and output structured data containing multiple preset fields; Multiple semantic analysis agents are used to perform semantic analysis on the structured data from multiple dimensions through parallel processing, including parsing and matching the work unit name field in the structured data using the organizational structure file; A central coordinator is used to schedule the concurrent execution of the multiple semantic analysis agents; The dynamic quantization rule engine is used to calculate the outputs of the multiple semantic analysis agents based on the pre-configured resume data quantization calculation logic, and generate the resume data quantization values ​​of each dimension and / or the overall resume data quantization value of the current person.

9. The system as described in claim 8, characterized in that, The document parsing intelligent agent is specifically used for: By calling the dynamic parsing service based on a large language model through the application programming interface, the appointment and removal table file is semantically understood and contextually reasoned to identify and extract data from the multiple preset fields. The extracted data is organized into a JavaScript object representation format that conforms to a predefined pattern; The structured data in the JavaScript object representation format is split according to the physical entity and stored in different database entities.

10. The system as described in claim 8, characterized in that, The plurality of semantic analysis agents include: An organizational structure parsing agent is used to match the work unit name field in the structured data to the organizational structure tree in the organizational structure file, and output the level and unique number. A professional line recognition intelligent agent is used to output one or more professional tags and functional types based on the job description text field in the structured data, combined with a predefined professional line dictionary and a large language model. A diversified experience judgment agent is used to determine, according to configurable rules, whether each work experience in the resume field of the structured data belongs to the special experience defined by policy. The grassroots experience analysis agent is used to determine whether the experience in the resume field meets the preset grassroots experience conditions by combining the organizational structure with the level to which the agent outputs. The qualification and honor analysis agent is used to standardize and map the professional and technical position field, academic institution field, and award and punishment field in the structured data and determine their level.

Images