User agreement-based labor service outsourcing intelligent management method and system

Abstract

The present application relates to the technical field of labor service outsourcing, in particular to a labor service outsourcing intelligent management method and system based on user agreement. The present application constructs a target knowledge graph and generates an updated correlation vector according to the correlation relationship, so that the selection of outsourcing personnel is more accurate, and the optimal outsourcing personnel can be dynamically matched based on the specific business needs of enterprises, thereby improving the adaptability of outsourcing personnel. Through the training of the weighted correlation vector and the standard matching prediction model, the system can better cope with different business needs, continuously optimize the outsourcing personnel matching recommendation effect, has strong self-learning and optimization ability, enhances the long-term adaptability of the system, and through the establishment of the knowledge graph, the updating of the correlation vector and the optimization of the matching prediction model, not only can meet the existing business needs, but also can be expanded and adjusted according to the specific requirements of different fields or industries, has strong system adaptability and flexibility.

Description

Technical Field

[0001] The present invention relates to the technical field of labor service outsourcing, and in particular to an intelligent management method and system for labor service outsourcing based on user agreements. Background Art

[0002] With the continuous optimization of corporate operating models, human resource outsourcing has become an important resource allocation strategy in modern enterprise management. By handing over non-core businesses to professionals from third-party suppliers, companies can effectively reduce operating costs, improve business execution efficiency, and concentrate resources on cultivating core competitiveness.

[0003] In the existing outsourcing model, companies usually establish cooperation with multiple suppliers. After the suppliers take over the business, they assign their business personnel to perform specific tasks. However, this model has significant limitations. The company's right to choose outsourced personnel is essentially limited by the supplier's assignment mechanism, making it difficult to dynamically match the best candidates based on specific business needs. Summary of the Invention

[0004] The main purpose of the present invention is to provide an intelligent management method for outsourcing employment services based on user agreements, aiming to solve the technical problems in the prior art.

[0005] The present invention proposes an intelligent management method for outsourcing labor services based on user agreements, comprising:

[0006] Acquire outsourced personnel data, and construct a target knowledge graph based on the outsourced personnel data, wherein the target knowledge graph includes a plurality of outsourced personnel nodes and a plurality of outsourced personnel associated nodes corresponding to each outsourced personnel node;

[0007] Acquire a corresponding association relationship according to each of the outsourced personnel nodes and the outsourced personnel association nodes;

[0008] Generate an updated association vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node according to the target knowledge graph;

[0009] Determine the actual weight between each outsourced personnel node and multiple outsourced personnel associated nodes according to the association relationship, and add each of the actual weights to the corresponding updated association vector to obtain a corresponding weighted association vector;

[0010] Training a preset matching prediction model according to the plurality of weighted association vectors to obtain a standard matching prediction model;

[0011] A query request is obtained and input into a job matching prediction model to make an outsourced personnel recommendation.

[0012] Preferably, before the step of generating, according to the target knowledge graph, an updated association vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node, the step includes:

[0013] Obtaining, from the target knowledge graph, M initial node groups corresponding to each outsourced personnel node, wherein the Nth initial node group includes N-1 feature nodes associated with the outsourced personnel node, where N is a positive integer and N∈[1,M];

[0014] Obtaining a specified time, and removing characteristic nodes that do not fall within the specified time from the M initial node groups corresponding to the outsourced personnel nodes, to obtain M updated node groups;

[0015] A feature vector is obtained according to the feature node, and the feature vectors of all feature nodes in each update node group are fused according to the outsourced personnel node to obtain a fused vector.

[0016] Preferably, the step of generating, according to the target knowledge graph, an updated association vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node comprises:

[0017] Acquire historical outsourced personnel data and current outsourced personnel data according to the outsourced personnel data, and construct a first target knowledge graph according to the historical outsourced personnel data;

[0018] Generating a first association fusion vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node according to the first target knowledge graph;

[0019] Constructing a second target knowledge graph based on the current outsourced personnel data;

[0020] Generating a second association fusion vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node according to the second target knowledge graph;

[0021] The first associated fusion vector in the first target knowledge graph is replaced according to each of the second associated fusion vectors to obtain a corresponding updated associated vector.

[0022] Preferably, the step of replacing the first association fusion vector in the first target knowledge graph according to the second association fusion vector to obtain a corresponding updated association vector includes:

[0023] Obtain multiple first entities and second entities in the first target knowledge graph and the second target knowledge graph, and obtain corresponding first relationships and second relationships according to each of the first entities and the second entities;

[0024] Obtaining corresponding entity similarity according to each of the first entity and the second entity;

[0025] Obtaining a corresponding relationship similarity based on each of the first relationship and the second relationship, and obtaining a corresponding entity relevance based on each of the relationship similarities and entity similarities;

[0026] Obtaining a total similarity based on the correlations of the plurality of entities, and obtaining a difference between the first associated fusion vector and the second associated fusion vector based on the total similarity;

[0027] If the difference is greater than a preset threshold, marking the first associated fusion vector;

[0028] If the difference is less than a preset threshold, the first association fusion vector in the first target knowledge graph is replaced by the second association fusion vector to obtain an updated association vector.

[0029] Preferably, the step of inputting the query request into a job matching prediction model to recommend outsourced personnel includes:

[0030] Obtaining request information according to the query request, and extracting multiple demand elements in the request information, wherein the demand elements are mapped to outsourced personnel associated nodes in the target knowledge graph;

[0031] Inputting each of the demand elements into a work matching prediction model to obtain an actual associated fusion vector corresponding to each of the demand elements;

[0032] Obtaining a corresponding correlation matching degree according to each of the actual correlation fusion vectors and the updated correlation vector, and sorting the multiple correlation matching degrees to obtain a matching degree sorting table;

[0033] Selecting the outsourced personnel association node corresponding to the first-ranked association matching degree from the matching degree ranking table as the recommended association node;

[0034] The corresponding outsourcing personnel node is obtained according to the recommended associated node, and the user is recommended according to the outsourcing personnel node.

[0035] Preferably, after the step of acquiring the corresponding outsourcing personnel node according to the recommended associated node and recommending the user according to the outsourcing personnel node, the method further includes:

[0036] Obtaining the feedback result of the outsourced personnel node's recommendation to the user, and determining whether the feedback result meets the preset expected value;

[0037] If the feedback result reaches the preset expected value, the recommendation result is determined to be qualified;

[0038] If the feedback result does not meet the preset expected value, returning to the step of selecting the outsourced personnel associated node corresponding to the first associated matching degree from the matching degree ranking table;

[0039] The outsourced personnel association node corresponding to the second-ranked association matching degree is selected and re-used as the recommended association node until the feedback result reaches the preset expected value.

[0040] This application also provides an intelligent management system for outsourcing employment services based on user agreements, including:

[0041] A construction module, configured to obtain outsourced personnel data and construct a target knowledge graph based on the outsourced personnel data, wherein the target knowledge graph includes a plurality of outsourced personnel nodes and a plurality of outsourced personnel associated nodes corresponding to each outsourced personnel node;

[0042] An acquisition module, configured to acquire a corresponding association relationship according to each of the outsourced personnel nodes and the outsourced personnel association nodes;

[0043] A generation module, configured to generate, based on the target knowledge graph, an updated association vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node;

[0044] a determination module, configured to determine an actual weight between each outsourced personnel node and a plurality of outsourced personnel associated nodes according to the association relationship, and add each of the actual weights to a corresponding updated association vector to obtain a corresponding weighted association vector;

[0045] A training module, configured to train a preset matching prediction model based on a plurality of weighted association vectors to obtain a standard matching prediction model;

[0046] The recommendation module is used to obtain a query request and input the query request into the job matching prediction model to recommend outsourced personnel.

[0047] Preferably, the generating module comprises:

[0048] A first construction unit is configured to obtain historical outsourced personnel data and current outsourced personnel data according to the outsourced personnel data, and to construct a first target knowledge graph according to the historical outsourced personnel data;

[0049] A first generating unit is configured to generate, based on the first target knowledge graph, a first association fusion vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node;

[0050] A second construction unit is configured to construct a second target knowledge graph based on the current outsourced personnel data;

[0051] A second generating unit is configured to generate, based on the second target knowledge graph, a second association fusion vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node;

[0052] A replacement unit is used to replace the first associated fusion vector in the first target knowledge graph according to each of the second associated fusion vectors to obtain a corresponding updated associated vector.

[0053] The present invention also provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned intelligent management method for employment service outsourcing based on user agreements are implemented.

[0054] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned intelligent management method for employment service outsourcing based on user agreements.

[0055] The beneficial effects of the present invention are as follows: the present invention makes the selection of outsourced personnel more accurate by constructing a target knowledge graph and generating an updated association vector based on the association relationship, and can dynamically match the optimal outsourced personnel based on the specific business needs of the enterprise, thereby improving the adaptability of the outsourced personnel and avoiding the limitations of the traditional supplier assignment mechanism. Through the intelligent management system, the enterprise can select outsourced personnel more flexibly and efficiently, avoiding dependence on the supplier's single assignment mechanism, so as to better adjust personnel allocation according to business needs, improve resource utilization efficiency, directly control the selection of outsourced personnel, reduce uncertainty and information asymmetry in outsourcing services, ensure a high degree of fit between outsourced personnel and enterprise needs, and thus improve the overall quality of outsourcing services. Through the training of weighted association vectors and standard matching prediction models, the system can better It can better respond to different business needs, continuously optimize the matching and recommendation effects of outsourced personnel, has strong self-learning and optimization capabilities, and enhances the long-term adaptability of the system. The automated outsourced personnel recommendation system greatly reduces the time and energy costs of enterprises in selecting outsourced personnel and improves decision-making efficiency. With the assistance of intelligent algorithms, enterprises can make more scientific and reasonable outsourced personnel selection decisions and promote the digital transformation of enterprise management methods. By establishing knowledge graphs, updating association vectors and optimizing matching prediction models, it can not only meet existing business needs, but also expand and adjust according to the specific requirements of different fields or industries. It has strong system adaptability and flexibility, effectively solving the limitations of the supplier assignment mechanism in the traditional outsourcing model, and improving the efficiency and quality of outsourcing management through intelligent management and precise personnel matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 Schematic diagram of a method flow according to an embodiment of the present invention.

[0057] Figure 2 FIG. 1 is a schematic diagram of a system structure according to an embodiment of the present invention.

[0058] Figure 3 This is a schematic diagram of the internal structure of a computer device according to an embodiment of the present application.

[0059] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0060] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0061] like Figure 1 As shown, the present application provides an intelligent management method for outsourcing employment services based on user agreements, including:

[0062] S1. Obtain outsourced personnel data, and construct a target knowledge graph based on the outsourced personnel data, wherein the target knowledge graph includes multiple outsourced personnel nodes and multiple outsourced personnel associated nodes corresponding to each outsourced personnel node;

[0063] S2. Acquire a corresponding association relationship based on each of the outsourced personnel nodes and the outsourced personnel association nodes;

[0064] S3. Generate updated association vectors between multiple outsourced personnel-associated nodes that have an association relationship with the outsourced personnel node according to the target knowledge graph;

[0065] S4. Determine the actual weight between each outsourced personnel node and multiple outsourced personnel associated nodes according to the association relationship, and add each of the actual weights to the corresponding updated association vector to obtain a corresponding weighted association vector;

[0066] S5. Training a preset matching prediction model according to the plurality of weighted association vectors to obtain a standard matching prediction model;

[0067] S6. Obtain a query request, and input the query request into a job matching prediction model to recommend outsourced personnel.

[0068] As described in the above steps S1-S6, the present invention obtains the outsourced personnel data and constructs a target knowledge graph based on the outsourced personnel data, wherein the target knowledge graph includes multiple outsourced personnel nodes and multiple outsourced personnel associated nodes corresponding to each outsourced personnel node. By obtaining the outsourced personnel data, multi-dimensional information related to the outsourced personnel, such as skills, work experience, performance, etc., can be obtained. By collecting and analyzing various types of data on outsourced personnel, the accuracy and practicality of the knowledge graph can be guaranteed. By comprehensively collecting data, a basis is provided for subsequent intelligent matching, avoiding a single, static personnel selection method. Acquiring outsourced personnel data is the basic step of the entire process, providing a basis for subsequent knowledge graph construction and generation of matching prediction models. Providing necessary data support, building a target knowledge graph can associate outsourced personnel with their associated nodes, which can more clearly present the relationship between outsourced personnel and business needs. By establishing the relationship between outsourced personnel and their associated nodes, the multi-dimensional characteristics of outsourced personnel (such as skills, experience, etc.) can be revealed, thereby providing valuable information for intelligent recommendation and dynamic matching. The relationship between outsourced personnel nodes and associated nodes can fully reflect the capabilities and work background of outsourced personnel, avoiding the limitations of traditional static screening methods. Most existing technologies rely on manual judgment or simple empirical rules and cannot systematically match outsourced personnel with business needs. Through the construction of knowledge graphs, automated and data-driven matching can be achieved;

[0069] By obtaining the corresponding association relationship between each outsourced personnel node and the outsourced personnel associated node, the relationship between the outsourced personnel node and the associated node can be accurately depicted. Through the detailed association relationship, it is possible to identify which outsourced personnel are more suitable for specific task requirements. The association relationship can be automatically extracted through data mining or machine learning algorithms without manual intervention, reducing human errors and biases. By clarifying the association relationship, such mismatches can be eliminated and the accuracy of selection can be improved. The acquisition of association relationships is closely related to the construction of the knowledge graph, which provides an important basis for the subsequent update of association vectors and the generation of weighted association vectors.

[0070] By generating updated association vectors between multiple outsourced personnel associated nodes that have an association relationship with the outsourced personnel node through the target knowledge graph, the generated updated association vectors can dynamically reflect the changes in the relationship between the outsourced personnel and the associated nodes, making the selection of outsourced personnel more flexible and accurate. In order to achieve more intelligent matching, the association relationship at different time points may change. By generating updated association vectors, such changes can be reflected in real time to ensure the timeliness and accuracy of the recommendation results. The generated updated association vectors can be refined to the specific matching situation of each outsourced personnel node, optimizing resource allocation. Unlike the static matching method in the prior art, the present invention can more flexibly and accurately match personnel according to actual needs by dynamically updating the association relationship vectors. The generated updated association vectors provide a basis for dynamic adjustment for the subsequent generation of weighted association vectors.

[0071] The actual weight between each outsourced personnel node and multiple outsourced personnel associated nodes is determined through the association relationship, and each actual weight is added to the corresponding updated association vector to obtain the corresponding weighted association vector. By calculating the actual weight and generating the weighted association vector, the actual degree of association between each outsourced personnel node and its associated nodes can be accurately reflected, avoiding the "homogenization" problem. The calculation of the weight can weight the association relationship according to the actual ability, experience and other factors of the outsourced personnel, so that the matching results are more in line with actual needs. By assigning different weights, the matching algorithm can be optimized to select the most suitable outsourced personnel, avoiding simple recommendations based on equal weights. Traditional outsourced personnel selection usually ignores individual differences and fails to accurately reflect the actual ability of each outsourced personnel. Through the weighting mechanism, more detailed personalized recommendations can be achieved. The generation of the weighted association vector is a further optimization of the updated association vector, making the selection of outsourced personnel more accurate.

[0072] The preset matching prediction model is trained through multiple weighted association vectors to obtain a standard matching prediction model. The trained matching prediction model can make automatic and accurate recommendations based on the relationship between outsourced personnel and tasks. By training the matching prediction model, historical data, association relationships and weighted vectors can be effectively utilized to build an intelligent matching system with strong predictive capabilities. Through training with machine learning algorithms, effective matching patterns can be automatically extracted from data without relying on manual experience. The selection of outsourced personnel in the existing technology usually relies on manual rules or static models, lacking adaptability and intelligence. The present invention can greatly improve the accuracy and efficiency of selection through the intelligent prediction model. By training the model, efficient matching with query requests can be achieved.

[0073] By obtaining a query request and inputting the query request into the job matching prediction model to recommend outsourced personnel, the system can dynamically respond to the needs of the enterprise and provide personalized and accurate outsourced personnel recommendations. By responding to the query request, real-time matching and recommendation can be achieved, and the selection of outsourced personnel can be adjusted in time according to changes in the enterprise's needs. The query request can reflect the current specific needs of the enterprise, and the outsourced personnel recommended by the system based on the model can ensure the efficient completion of the task. Traditional methods are often unable to dynamically recommend personnel based on specific business needs. The present invention can respond to changes in time through an intelligent model and provide more suitable personnel recommendations. The acquisition of the query request is closely related to the previous steps. Finally, the model generates recommendation results to achieve accurate matching. The present invention can achieve intelligent dynamic matching of outsourced personnel, thereby solving the problem of limited selection of outsourced personnel by enterprises in the existing outsourcing model, optimizing the selection mechanism of outsourced personnel, and making the matching between outsourced personnel and task requirements more accurate, flexible and efficient. Through the correlation modeling and updating based on the knowledge graph and the training of the intelligent matching prediction model, the present invention greatly improves the automation level of outsourced personnel selection and the accuracy of personalized recommendation, and has significant technical advantages and innovations.

[0074] In one embodiment, before step S3 of generating updated association vectors between a plurality of outsourced personnel-associated nodes that have an association relationship with an outsourced personnel node according to the target knowledge graph, the following steps are included:

[0075] S31. Obtain, from the target knowledge graph, M initial node groups corresponding to each outsourced personnel node, wherein the Nth initial node group includes N-1 feature nodes associated with the outsourced personnel node, where N is a positive integer and N∈[1,M];

[0076] S32: Obtain a specified time, and remove characteristic nodes that do not fall within the specified time from the M initial node groups corresponding to the outsourced personnel nodes to obtain M updated node groups;

[0077] S33 , acquiring a feature vector according to the feature node, and fusing the feature vectors of all feature nodes in each update node group according to the outsourced personnel node to obtain a fusion vector.

[0078] As described in the above steps S31-S33, the present invention obtains M initial node groups corresponding to the outsourced personnel node from the target knowledge graph through each outsourced personnel node, wherein the Nth initial node group includes N-1 feature nodes that have an association relationship with the outsourced personnel node. By extracting feature nodes related to the outsourced personnel from the knowledge graph, it is possible to accurately describe the capabilities, experience, historical tasks and other related information of the outsourced personnel, and construct a complete portrait of the outsourced personnel. In this way, the enterprise can make a selection based on the real background data of the outsourced personnel, thereby avoiding human bias or insufficient judgment. In the traditional outsourcing model, the enterprise's choice of outsourced personnel is limited to the assignment of suppliers, which is often difficult to meet the dynamic nature of business needs. By dynamically obtaining the feature nodes of the outsourced personnel, it can provide for the subsequent matching process. It provides more detailed information, enhances the flexibility and adaptability of selection, and when obtaining M initial node groups, the feature nodes in each initial node group can cover the correlation between outsourced personnel and tasks, and can provide accurate feature input for selection and matching. The node association based on the graph can more comprehensively and dynamically evaluate the adaptability of outsourced personnel. In the existing technology, the selection of outsourced personnel mainly relies on the static assignment of suppliers and lacks a dynamic evaluation mechanism based on the knowledge graph. The present invention can highly match the selection of outsourced personnel with specific business needs, thereby improving the accuracy and matching of selection, laying the foundation for subsequent time screening and feature vector fusion, providing preliminary feature data of outsourced personnel, and providing basic information of outsourced personnel through associated feature nodes, providing key input for data update and fusion in subsequent steps;

[0079] By obtaining a specified time and removing the feature nodes that do not fall within the specified time from the M initial node groups corresponding to the outsourced personnel nodes, M updated node groups are obtained. By setting a time filtering condition, it is possible to ensure that the feature information of each outsourced personnel is the latest and most relevant within a specific time range, thereby avoiding outdated or irrelevant information from interfering with decision-making. The capabilities and adaptability of outsourced personnel may change over time. For example, outsourced personnel may participate in different projects, accumulate new experience, or undergo position changes. The present invention can ensure that the system selects the most relevant and up-to-date outsourced personnel information that is currently required, thereby improving the accuracy of matching. The time screening parameter ensures that the selected feature nodes are timely by limiting the time range of the feature nodes, effectively avoiding outdated evaluations of the outsourced personnel's capabilities and ensuring a more accurate match between business needs and outsourced personnel. Traditional outsourced personnel evaluations usually ignore the time factor, which may result in the selected information no longer being suitable for current needs. By introducing a time filtering mechanism, this defect is avoided, making the selection process more accurate. The initial node group obtained in the first step is screened to obtain the latest node data that is relevant to the specific time period, providing high-quality data input for subsequent feature vector processing.

[0080] Acquiring feature vectors through feature nodes can convert the actual meaning of each feature node into a data format that can be processed by a computer, and then perform subsequent data analysis and fusion. Feature vectors can accurately represent the capabilities, experience and other characteristics of outsourced personnel, and have highly digitized and quantitative characteristics, which is convenient for subsequent analysis. Converting feature nodes into feature vectors is the basis for subsequent machine learning and data analysis, and can convert complex node data into processable numerical information, thereby facilitating the system to perform effective fusion and matching. In the present invention, the feature information of outsourced personnel can be converted into numerical vectors, which is convenient for the system to calculate and match, can effectively reduce the deviation of manual operation, and more accurately quantify the characteristics of outsourced personnel. Traditional outsourced personnel evaluation relies on manual or static rule judgment, and lacks in-depth quantitative analysis of outsourced personnel characteristics. The present invention can provide higher accuracy and more comprehensive data support for the matching of outsourced personnel through feature vector quantization processing;

[0081] According to the outsourced personnel node, the feature vectors of all feature nodes in each update node group are fused to obtain a fusion vector. The fusion of feature vectors can integrate the information of multiple feature nodes to obtain a more representative vector, which fully expresses the comprehensive ability of the outsourced personnel. This fusion process can improve the system's evaluation accuracy of outsourced personnel and ensure that multi-dimensional information is comprehensively considered. Different feature nodes may reflect the ability of outsourced personnel from different angles. By fusing multiple feature vectors, a more comprehensive and comprehensive vector can be obtained, which enhances the accuracy of matching and multi-dimensional considerations. By fusing the feature vectors of different feature nodes, the performance and adaptability of outsourced personnel in different dimensions can be better captured, avoiding the limitations that may be brought about by the evaluation of a single dimension. In the existing technology, it may only rely on The evaluation of outsourced personnel relies on a single dimension or a simple scoring method, which lacks the ability to comprehensively consider multi-dimensional factors. The present invention can conduct a more comprehensive evaluation of outsourced personnel through fusion processing, thereby improving the matching effect. This method is the core part of the entire process. By fusing feature vectors, the comprehensive evaluation results of outsourced personnel are finally obtained. The work results of the first three steps (initial node acquisition, time screening, feature vector conversion) are integrated to provide core data support for the final outsourced personnel matching. By gradually analyzing the beneficial effects of the above steps, the present invention can effectively solve the limitations of the existing outsourcing model and improve the flexibility, accuracy and adaptability of outsourced personnel selection by dynamically acquiring the feature nodes of outsourced personnel, introducing a time screening mechanism, feature vector quantization and fusion processing.

[0082] In one embodiment, step S3 of generating updated association vectors between a plurality of outsourced personnel-associated nodes that have an association relationship with an outsourced personnel node according to the target knowledge graph includes:

[0083] S34. Acquire historical outsourced personnel data and current outsourced personnel data based on the outsourced personnel data, and construct a first target knowledge graph based on the historical outsourced personnel data;

[0084] S35. Generate a first association fusion vector between a plurality of the outsourced personnel-associated nodes that have an association relationship with the outsourced personnel node according to the first target knowledge graph;

[0085] S36. Construct a second target knowledge graph based on the current outsourced personnel data;

[0086] S37. Generate a second association fusion vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node according to the second target knowledge graph;

[0087] S38. Replace the first associated fusion vector in the first target knowledge graph according to each of the second associated fusion vectors to obtain a corresponding updated associated vector.

[0088] As described in the above steps S34-S38, the present invention obtains historical outsourced personnel data and current outsourced personnel data through outsourced personnel data. By obtaining historical outsourced personnel data and current outsourced personnel data, the system can perform more comprehensive data analysis and personnel matching, not only providing multi-dimensional information about outsourced personnel, but also allowing a comprehensive evaluation of their historical performance, which helps to mine more valuable outsourced personnel data patterns. Through the combination of historical data and current data, the system can better understand the outsourced personnel's work ability, historical task performance, task completion status, work efficiency and other aspects, and provide a more accurate basis for personnel selection. The existing technology often relies only on the current outsourced personnel information provided by the outsourcing company, but ignores the outsourced personnel's historical data. Historical data can provide more in-depth analysis, such as judging the outsourced personnel's performance and adaptability in similar tasks, etc., so as to avoid enterprises making wrong decisions due to incomplete or inaccurate information. This step provides a data source basis, provides data support for the subsequent construction of the target knowledge graph, generation of associated fusion vectors and updating of associated vectors, and is the data input link in the entire process;

[0089] The first target knowledge graph is constructed based on historical outsourced personnel data. The construction of the first target knowledge graph converts the data of outsourced personnel into structured information, which can reveal the correlation between outsourced personnel. Through the first target knowledge graph, enterprises can more clearly see the similarities and complementarities between different outsourced personnel, which plays a key role in subsequent personnel matching and optimization. The first target knowledge graph can construct the relationships and characteristics between outsourced personnel based on historical data. The construction of the first target knowledge graph provides basic data support for subsequent data processing and analysis, and can discover the potential capabilities and matching value of personnel through the relationship between nodes. Existing technologies usually lack in-depth correlation analysis of outsourced personnel and cannot display the multi-dimensional relationship between outsourced personnel through graphs. Compared with traditional personnel screening methods, graph methods can more intuitively display the potential of outsourced personnel and provide enterprises with more selection basis. This step is the core of the entire process. The construction of the first target knowledge graph provides a structured information foundation for subsequent steps such as generating and updating correlation fusion vectors, ensuring the accuracy and effectiveness of subsequent steps.

[0090] A first association fusion vector is generated between multiple outsourced personnel-related nodes that have an association relationship with the outsourced personnel node through the first target knowledge graph. By generating the first association fusion vector, the relationship between the outsourced personnel and other associated nodes (such as other outsourced personnel or tasks) can be quantified to form a vector expression. This vectorization method makes subsequent calculations and analysis more efficient and accurate. The first association fusion vector can focus on the actual working relationship, skill requirements, task coordination, etc. between outsourced personnel, avoiding simple manual screening, reducing subjective bias, and improving matching efficiency. Traditional technologies mostly rely on manual evaluation or screening based on simple conditions. The present invention avoids manual errors through the generation of association vectors, improves the accuracy and objectivity of outsourced personnel selection, and generates the first association fusion vector to dynamically match outsourced personnel by comparing and updating association vectors in subsequent steps, so that the selection of outsourced personnel is more in line with the actual needs of the enterprise.

[0091] The second target knowledge graph is constructed through the current outsourced personnel data. The second target knowledge graph is based on the current outsourced personnel data and can reflect the performance of the outsourced personnel in the current environment, task matching, etc. The second target knowledge graph is based on the current outsourced personnel data and can reflect the changes and adaptability of the outsourced personnel in real time. By constructing the second target knowledge graph, the enterprise can obtain an accurate response to real-time task requirements. The existing technology usually lacks dynamic analysis based on the current outsourced personnel performance. The present invention can dynamically respond to the actual situation of the outsourced personnel and provide real-time optimization support by constructing the second target knowledge graph. The construction of the second target knowledge graph echoes the first target knowledge graph, providing a new basis for the subsequent update of the associated fusion vector. By comparing historical and current data, the outsourced personnel selection strategy can be dynamically adjusted to improve the matching accuracy.

[0092] A second association fusion vector is generated between multiple outsourced personnel association nodes that have an association relationship with the outsourced personnel node through the second target knowledge graph. The second association fusion vector is generated based on the second target knowledge graph, which means that the vector can accurately reflect the adaptability and potential capabilities of the current outsourced personnel in the actual task, which helps to adjust personnel matching more dynamically. Through the second association fusion vector, the enterprise can obtain personnel recommendations that are more in line with the current task requirements. The second association fusion vector can reflect the actual work association between personnel, avoiding the lag problem that may occur in traditional methods. Traditional outsourced personnel matching methods are usually unable to respond to changes in task requirements in real time. The present invention can perform accurate matching in a short time by generating a second association fusion vector, thereby improving the utilization efficiency of outsourced personnel. The generation of the second association fusion vector directly provides support for updating the vector in the first target knowledge graph. By replacing the vector in the first target knowledge graph with the second association fusion vector, the update and dynamic adjustment of the knowledge graph are ensured, making subsequent personnel matching more accurate and flexible.

[0093] By replacing the first associated fusion vector in the first target knowledge graph with each second associated fusion vector, a corresponding updated associated vector is obtained. By replacing the first associated fusion vector in the first target knowledge graph with the second associated fusion vector, dynamic updating of the knowledge graph can be achieved, thereby improving the real-time and accuracy of personnel matching. The updated graph can more accurately reflect the current status and task requirements of the outsourced personnel. Replacing the updated associated vector enables the knowledge graph to continuously optimize itself and improve the flexibility and accuracy of outsourced personnel selection. The graphs in traditional methods are usually static and cannot be quickly updated according to the latest data. The present invention dynamically updates the associated vectors so that the knowledge graph can flexibly respond to changes in outsourced personnel and ensure that the matching results are more in line with real-time needs. This step is a comprehensive application and optimization of the aforementioned steps, ensuring that the knowledge graph finally generated can accurately reflect the matching situation between current task requirements and outsourced personnel, and provide enterprises with the best personnel selection strategy.

[0094] In one embodiment, the step S38 of replacing the first association fusion vector in the first target knowledge graph according to the second association fusion vector to obtain a corresponding updated association vector includes:

[0095] S381, obtaining multiple first entities and second entities in the first target knowledge graph and the second target knowledge graph, and obtaining corresponding first relationships and second relationships according to each of the first entities and the second entities;

[0096] S382: Calculate corresponding entity similarity based on the cosine similarity between each of the first entity and the second entity;

[0097] S383: Calculate the corresponding relationship similarity based on each of the first relationship and the second relationship by using cosine similarity, and calculate the corresponding entity relevance based on the weighted sum of each of the relationship similarities and the entity similarity;

[0098] S384. Obtain a total similarity based on the correlations of the plurality of entities, and obtain a difference between the first associated fusion vector and the second associated fusion vector based on the total similarity (the sum of the difference and the total similarity is 1);

[0099] If the difference is greater than a preset threshold, marking the first associated fusion vector;

[0100] If the difference is less than a preset threshold, the first association fusion vector in the first target knowledge graph is replaced by the second association fusion vector to obtain an updated association vector.

[0101] As described in the above steps S381-S384, the present invention obtains multiple first entities and second entities in the first target knowledge graph and the second target knowledge graph. By obtaining multiple entities, it is possible to comprehensively and systematically compare and analyze the information in different target knowledge graphs, providing richer background information. The entities in different graphs can reflect the data characteristics of different fields or different systems, providing a multi-dimensional data source for subsequent analysis, and ensuring that meaningful entity data is obtained in the multi-dimensional graph, laying the foundation for subsequent similarity calculation and entity correlation analysis. By selecting multiple entities, the comprehensiveness of the calculation can be guaranteed, avoiding the omission of important data, and ensuring the depth and breadth of the comparison. The existing technology often focuses on entities in a single graph, lacks cross-graph entity comparison, and is difficult to fully reflect the correlation between different fields. The present method makes up for the shortcomings of the existing technology through cross-graph entity selection. The acquisition of entities is the basis of the entire process. The relationship similarity, entity similarity and correlation calculations in the subsequent steps are based on this. By ensuring the accurate acquisition of entities, the accuracy of subsequent analysis can be improved;

[0102] And obtain the corresponding first relationship and second relationship based on each first entity and second entity. By clarifying the relationship between the first entity and the second entity, their internal connection can be deeply explored, which is helpful for accurate similarity analysis. By extracting the upstream and downstream relationships, dependencies, etc. between entities, it is helpful for detailed comparison and analysis in subsequent steps, ensuring the integrity of relationship information. It is the link between entities. By obtaining the corresponding relationship, the connection between entities can be deeply understood, and more detailed parameters are provided for similarity calculation. Obtaining the corresponding relationship helps to avoid incorrect matching or overly general similarity evaluation and improve the accuracy of calculation. In traditional knowledge graph analysis, subtle relationship differences between entities may be ignored, resulting in inaccurate analysis results. This method improves the comparability and analysis depth of graph data by clarifying the relationship. It is closely related to the first step. The obtained relationship data is the basis for subsequent similarity calculation and correlation evaluation. The relationship information can further refine the similarity calculation process between entities.

[0103] By obtaining the corresponding entity similarity of each first entity and second entity, and through quantitative entity similarity evaluation, a numerical basis is provided for subsequent correlation analysis, avoiding subjective errors. The evaluation of entity similarity can reveal the potential connections between different fields and provide data support for optimized decision-making. The similarity evaluation can intuitively reflect the degree of similarity between entities and provide a quantitative basis for correlation judgment. By introducing the calculation of similarity, manual intervention and uncertainty of judgment are avoided. Entity similarity provides a measurable standard that can systematically compare different entities and reflect the potential correlation in the graph. Traditional methods often ignore subtle similarity evaluations across graphs and fields, which can easily lead to underestimation or overestimation of correlation. Through this step, the similarity between different entities can be calculated and compared more accurately. This step is closely related to the previous entity acquisition and relationship analysis, and provides an important basis for subsequent correlation evaluation and updating of fusion vectors.

[0104] By obtaining the corresponding relationship similarity through each first relationship and second relationship, the evaluation of relationship similarity makes the analysis more detailed, can quantify the strength of similar relationships in different graphs, and further optimize the effect of knowledge fusion. Relationship similarity can help determine whether two relationships are similar in function and structure, and is a key factor in relationship matching. By calculating relationship similarity, the subsequent knowledge fusion process is ensured to be more accurate. The quantification of similarity can avoid ambiguity and improve the accuracy of the system when dealing with complex relationships. Existing methods usually ignore the impact of relationship matching on the overall analysis. This method improves the efficiency and accuracy of graph integration by accurately evaluating relationship similarity, works closely with entity similarity analysis, further optimizes similarity matching through relationship similarity, and provides a basis for subsequent fusion vector adjustment and difference calculation.

[0105] The corresponding entity relevance is obtained based on each relationship similarity and entity similarity. Combining relationship similarity with entity similarity can comprehensively consider the influence of the two dimensions and more comprehensively evaluate the relevance between entities. By combining the similarity of entities and relationships, the association between entities in the graph can be more accurately described, thereby improving the quality of subsequent processing. Combining the similarity of multiple dimensions can comprehensively consider the impact of different factors on entity relevance and improve the comprehensiveness of analysis. Existing technologies may process entities and relationships independently, resulting in one-sided analysis. The combined processing of this method ensures full consideration of multi-dimensional data, improves the accuracy and adaptability of the system, and combines the key step of combining the similarity of the previous two dimensions (entity and relationship), providing a basis for subsequent total similarity calculation;

[0106] Obtaining total similarity through the correlation of multiple entities. The calculation of total similarity can comprehensively evaluate the similarity between multiple entities and provide a global perspective for further decision-making. By summarizing the correlation of multiple entities, we can avoid the one-sidedness of local analysis and ensure the accuracy of overall judgment. By summarizing the correlation of multiple entities, we can obtain a more comprehensive similarity measurement and avoid the deviation caused by isolated similarity calculation. Traditional methods often focus on individual comparisons and lack a unified overall evaluation. This method can more effectively integrate information from different sources. This method is an extension of the previous entity correlation calculation and uses a summary method to improve the integrity and effectiveness of the total similarity calculation.

[0107] The difference between the first associated fusion vector and the second associated fusion vector is obtained based on the total similarity. By calculating the difference, the difference between the two fusion vectors can be quantified, providing a basis for further decision-making. The difference calculation can reveal the difference between the two fusion vectors, helping to determine whether the fusion vector needs to be replaced or adjusted, improving the flexibility of the system. It can provide an intuitive difference assessment and avoid blindly replacing or marking vectors. The existing technology lacks a quantitative assessment of vector differences, making it impossible to accurately determine when to update the vector. This step can provide a reasonable basis for adjustment. The difference calculation relies on the total similarity assessment, ensuring that the basis for replacement or marking has sufficient logical support.

[0108] If the difference is greater than the preset threshold, the first associated fusion vector is marked. If the difference is less than the preset threshold, the second associated fusion vector replaces the first associated fusion vector in the first target knowledge graph to obtain an updated associated vector. Intelligent judgment is made through the preset threshold, and the vector fusion result can be automatically adjusted to reduce the need for manual intervention. Through threshold setting, it is possible to intelligently judge whether to replace or mark the vector based on the difference, thereby improving the dynamic adaptability and accuracy of the system. The preset threshold provides a flexible adjustment mechanism that can be customized according to actual needs, enhancing the versatility and adjustability of the method. Existing technologies usually lack automatic adjustment mechanisms and rely on manual decision-making. This step improves the automation and adaptability of the system through intelligent processing. Based on the difference judgment, it is further decided whether to replace or mark. It is closely related to the previous similarity calculation and total similarity evaluation, providing the system with dynamic update capabilities.

[0109] In one embodiment, the step S6 of inputting the query request into the job matching prediction model to recommend outsourced personnel includes:

[0110] S61. Obtain request information according to the query request, and extract multiple demand elements in the request information, wherein the demand elements are mapped to outsourced personnel associated nodes in the target knowledge graph;

[0111] S62: Input each demand factor into a work matching prediction model to obtain an actual associated fusion vector corresponding to each demand factor;

[0112] S63: Calculate the corresponding correlation matching degree based on each of the actual correlation fusion vectors and the updated correlation vector by using cosine similarity, and sort the multiple correlation matching degrees to obtain a matching degree sorting table;

[0113] S64: Select the outsourced personnel associated node corresponding to the first associated matching degree from the matching degree ranking table as the recommended associated node;

[0114] S65: Obtain a corresponding outsourcing personnel node according to the recommended associated node, and recommend the user according to the outsourcing personnel node.

[0115] As described in the above steps S61-S65, the present invention obtains request information through query requests. By obtaining the request information, it can ensure that the system obtains the specific needs of users or enterprises in real time, and provide accurate basic data for the analysis of subsequent steps. By querying the request information, it can flexibly adapt to different demand scenarios, avoid the single preset data input method, improve the universality and scalability of the system, and is closely related to the subsequent demand factor extraction and matching model input, ensuring that the subsequent demand factors can be effectively mapped and processed based on accurate data, and extracting multiple demand factors in the request information, wherein the demand factors are mapped to the outsourced personnel association node in the target knowledge graph. Extracting multiple demand factors in the request information helps to understand user needs more carefully, not just limited to a single demand, and improves the accuracy of demand analysis. User needs often include multiple aspects, and extracting multiple demand elements It can ensure that all dimensions of the demand are fully captured to avoid ignoring certain details. Under the existing outsourcing model, the demand is usually simplified to a single dimension, which cannot effectively handle complex and multi-dimensional demands. The present invention can better match suitable candidates by extracting multiple demand factors. The extracted demand factors will be mapped to the outsourced personnel nodes in the target knowledge graph and become the input of the subsequent work matching prediction model, ensuring that the subsequent steps can be effectively matched based on comprehensive demand. By mapping the demand factors to the outsourced personnel associated nodes in the target knowledge graph, intelligent matching can be achieved, and the associated relationships in the graph can be used to efficiently screen out personnel that match the demand. The knowledge graph can effectively integrate various business knowledge and personnel information, making the matching between demand and outsourced personnel more accurate. After being mapped to the knowledge graph, the demand factors provide data support for the subsequent work matching prediction model and promote the subsequent associated matching degree calculation;

[0116] By inputting each demand factor into the job matching prediction model, the actual associated fusion vector corresponding to each demand factor is obtained. The job matching prediction model can be used to dynamically judge the matching degree between demand factors and outsourced personnel based on historical data and algorithm prediction, thereby improving the accuracy and reliability of matching. With the support of the prediction model, manual intervention can be reduced and matching efficiency can be improved. The prediction model can optimize the matching strategy according to the ever-changing needs and market conditions. Existing technologies usually rely on static rules or manual judgment and cannot be dynamically adjusted according to changing needs. With the help of algorithms, the job matching prediction model significantly improves flexibility and adaptability. The output of the job matching prediction model is The actual correlation fusion vector provides basic data support for subsequent correlation matching calculations. The generated actual correlation fusion vector is a synthesis of multi-dimensional data and can accurately reflect the matching between demand factors and outsourced personnel, avoiding the limitations of a single dimension. By fusing multiple feature information, it can more comprehensively consider the matching relationship between demand and outsourced personnel, improving the accuracy of matching results. Traditional methods are unable to perform multi-dimensional fusion matching and usually rely on manual screening or simple rules. The actual correlation fusion vector can more comprehensively express the matching degree between demand and personnel. The actual correlation fusion vector is closely related to the subsequent correlation matching calculation and is an important input data for determining the final matching degree.

[0117] The corresponding correlation matching degree is obtained through each actual correlation fusion vector and the updated correlation vector. By calculating the correlation matching degree, the fit between the demand factors and the outsourced personnel can be quantified, providing a scientific basis for decision-making and avoiding deviations caused by subjective judgment. By introducing the updated correlation vector, the matching strategy can be dynamically adjusted and optimized according to actual conditions, making the matching process more flexible and accurate. Existing technologies often use static matching algorithms and lack dynamic adjustment capabilities. Introducing the updated correlation vector allows for flexible adjustment based on new requirements and external changes. The matching degree calculation provides a basis for subsequent sorting and recommendation steps, ensuring that the final recommended outsourced personnel best meet user needs.

[0118] Multiple correlation matching degrees are sorted to obtain a matching degree sorting table. By selecting the outsourced personnel correlation node corresponding to the correlation matching degree ranked first from the matching degree sorting table as the recommended correlation node, the most matching outsourced personnel can be quickly found through sorting, avoiding subjective bias in the manual screening process, making the recommendation results more objective and fair. Sorting can arrange all candidate personnel according to matching degree priority, thereby effectively screening out the best matching candidates and improving the efficiency and accuracy of recommendations. The selection of recommended correlation nodes ensures that the final recommendation is the outsourced personnel that best meets the user's needs, thereby improving the quality of outsourced personnel selection and reducing unnecessary matching and adjustments. By selecting the matching degree ranked first in the sorting, the relevance of the recommendation can be maximized, avoiding the selection of personnel with poor matching degrees. In the prior art, the selection of outsourced personnel mostly depends on the supplier's established rules or personnel assignments, while the present invention makes recommendations entirely based on needs and matching degrees, with higher degrees of freedom and accuracy. Recommending correlation nodes is a key step in the subsequent acquisition of outsourced personnel nodes, ensuring that the final selection of outsourced personnel meets actual needs.

[0119] By recommending associated nodes, corresponding outsourced personnel nodes are obtained, and recommendations are made to users based on the outsourced personnel nodes. Accurate personnel recommendations are provided through the finally selected outsourced personnel nodes to help users quickly find the most suitable outsourced personnel. The finally obtained outsourced personnel nodes can ensure that the personnel given by the recommendation system are the best choices based on the comprehensive matching algorithm, thereby improving the efficiency and quality of outsourced personnel. In traditional technologies, the selection of outsourced personnel is often constrained by suppliers, but the present invention can get rid of this limitation and provide more efficient personnel recommendations. The final outsourced personnel recommendation is the result of the entire process, reflecting the effectiveness and accuracy of the data processing, matching prediction, etc. in the previous steps. The technical solution of the present invention can effectively solve the limitations of the existing outsourcing model, provide a demand-based dynamic matching mechanism, avoid the limitations of supplier assignment in traditional methods, and improve the accuracy and flexibility of outsourced personnel selection.

[0120] In one embodiment, after the step S6 of acquiring the corresponding outsourcing personnel node according to the recommended associated node and recommending the user according to the outsourcing personnel node, the method further includes:

[0121] S66: Obtain the feedback result of the outsourced personnel node recommending the user, and determine whether the feedback result meets the preset expectation value;

[0122] If the feedback result reaches the preset expected value, the recommendation result is determined to be qualified;

[0123] If the feedback result does not meet the preset expected value, returning to the step of selecting the outsourced personnel associated node corresponding to the first associated matching degree from the matching degree ranking table;

[0124] S67: Select the outsourced personnel association node corresponding to the second-ranked association matching degree and re-use it as the recommended association node until the feedback result reaches the preset expected value.

[0125] As described in the above steps S66-S67, the present invention obtains the feedback results of the outsourced personnel node's recommendation to the user. By obtaining the feedback results of the outsourced personnel node's recommendation to the user, information on the degree of matching with the user's actual needs can be obtained in real time. This provides a data basis for subsequent adjustment of the recommendation results, ensuring that each recommendation can meet the user's needs as much as possible, thereby ensuring dynamic adjustment based on the actual feedback of the outsourced personnel node and the user. It is not necessary to rely on a static recommendation system, but to optimize the recommendation results based on the user's real feedback. The feedback results are dynamically obtained parameters and can be adjusted in real time according to different user groups and business types, reducing dependence on manual labor and improving the adaptability and flexibility of the system. The existing technology may only rely on system calculation or algorithm matching for recommendation, and may not take into account actual user feedback. The present invention can fundamentally improve the recommendation system by obtaining feedback in a timely manner, making it more in line with actual needs;

[0126] Determine whether the feedback result meets the preset expected value. If the feedback result meets the preset expected value, the recommendation result is determined to be up to standard. By comparing with the preset expected value, the quality of each recommendation is ensured to meet the standard, and meaningless recommendations that waste user and system resources are avoided. This ensures that the effect of the recommendation process is controllable. If the recommended feedback does not meet the preset expected value, the system will further optimize the recommendation to avoid user dissatisfaction or mismatching. The preset expected value is a standard set according to historical data and business needs to ensure that each recommendation is within a reasonable effect range and improve the accuracy of the recommendation. The existing technology may not have a clear effect judgment mechanism, and it is easy to have uneven recommendation effects. The present invention makes the recommendation more accurate by setting a clear expected value. The step directly determines whether to continue to execute the subsequent steps. Whether the feedback result meets the standard directly affects whether the recommendation needs to be optimized and adjusted;

[0127] If the feedback result does not meet the preset expected value, it returns to the step of selecting the outsourced personnel associated node corresponding to the first-ranked associated matching degree from the matching ranking table, and re-selects the outsourced personnel associated node corresponding to the second-ranked associated matching degree as the recommended associated node until the feedback result reaches the preset expected value. Through the fallback selection mechanism of the matching ranking table, the system can quickly make a second recommendation based on the outsourced personnel with higher matching degrees, thereby minimizing the occurrence of erroneous recommendations. When the feedback result does not meet the expected value, this design provides a fallback mechanism that can automatically select outsourced personnel with higher matching degrees for new recommendations, avoiding human intervention and improving matching efficiency. The matching ranking table ensures that each recommendation is based on the optimal matching degree ranking through multi-dimensional analysis of the outsourced personnel, and can continuously optimize in a data-driven manner. By dynamically adjusting the ranking of recommended personnel, it ensures that each adjustment is based on the best matching degree, continuously improving the accuracy and effect of recommendations, and by sequentially selecting outsourced personnel with high matching degree rankings, making the recommendation process flexible enough, and each adjustment can effectively improve the success rate of recommendations, avoiding too many invalid recommendations, and in each fallback process, the matching degree ranking table can be adjusted according to real-time feedback results, ensuring that each recommendation is more in line with user needs than the previous round of recommendations, avoiding manual repetitive operations. Compared with the existing outsourcing model, the present invention provides a more efficient dynamic personnel adjustment method, breaking through the limitation of outsourced personnel being fixedly assigned by suppliers, enhancing the flexibility of enterprises in personnel selection, and by gradually adjusting recommended personnel, maximizing the compliance with user needs, reducing the risk of mismatch, and improving overall work efficiency and user satisfaction.

[0128] like Figure 2 As shown, the present application also provides an intelligent management system for outsourcing employment services based on user agreements, including:

[0129] A construction module, configured to obtain outsourced personnel data and construct a target knowledge graph based on the outsourced personnel data, wherein the target knowledge graph includes a plurality of outsourced personnel nodes and a plurality of outsourced personnel associated nodes corresponding to each outsourced personnel node;

[0130] An acquisition module, configured to acquire a corresponding association relationship according to each of the outsourced personnel nodes and the outsourced personnel association nodes;

[0131] A generation module, configured to generate, based on the target knowledge graph, an updated association vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node;

[0132] a determination module, configured to determine an actual weight between each outsourced personnel node and a plurality of outsourced personnel associated nodes according to the association relationship, and add each of the actual weights to a corresponding updated association vector to obtain a corresponding weighted association vector;

[0133] A training module, configured to train a preset matching prediction model based on a plurality of weighted association vectors to obtain a standard matching prediction model;

[0134] The recommendation module is used to obtain a query request and input the query request into the job matching prediction model to recommend outsourced personnel.

[0135] In one embodiment, the generating module includes:

[0136] A first construction unit is configured to obtain historical outsourced personnel data and current outsourced personnel data according to the outsourced personnel data, and to construct a first target knowledge graph according to the historical outsourced personnel data;

[0137] A first generating unit is configured to generate, based on the first target knowledge graph, a first association fusion vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node;

[0138] A second construction unit is configured to construct a second target knowledge graph based on the current outsourced personnel data;

[0139] A second generating unit is configured to generate, based on the second target knowledge graph, a second association fusion vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node;

[0140] A replacement unit is used to replace the first associated fusion vector in the first target knowledge graph according to each of the second associated fusion vectors to obtain a corresponding updated associated vector.

[0141] It should be noted that each module and unit in the intelligent management system for outsourcing employment services based on user agreements corresponds one-to-one to the steps in the intelligent management method for outsourcing employment services based on user agreements.

[0142] like Figure 3 As shown, the present application also provides a computer device, which can be a server, and its internal structure can be as shown in FIG. Figure 3As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer design is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store all data required for the process of the intelligent management method for outsourcing of employment services based on user agreements. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the intelligent management method for outsourcing of employment services based on user agreements is implemented.

[0143] Those skilled in the art will understand that Figure 3 The structure shown in is merely a block diagram of a portion of the structure related to the present application solution and does not constitute a limitation on the computer device to which the present application solution is applied.

[0144] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, any one of the above-mentioned intelligent management methods for outsourcing employment services based on user agreements is implemented.

[0145] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided in this application and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct RAM bus dynamic RAM (DRDRAM), and RAM bus dynamic RAM (RDRAM).

[0146] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, apparatus, article, or method comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or method. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, apparatus, article, or method comprising the element.

[0147] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the description and drawings of the present invention, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. An intelligent management method for outsourcing labor services based on user agreements, characterized in that: include: Acquire outsourced personnel data, and construct a target knowledge graph based on the outsourced personnel data, wherein the target knowledge graph includes a plurality of outsourced personnel nodes and a plurality of outsourced personnel associated nodes corresponding to each outsourced personnel node; Acquire a corresponding association relationship according to each of the outsourced personnel nodes and the outsourced personnel association nodes; Generate updated association vectors between multiple outsourced personnel associated nodes that have an association relationship with the outsourced personnel node according to the target knowledge graph, including: obtaining historical outsourced personnel data and current outsourced personnel data according to the outsourced personnel data, and constructing a first target knowledge graph according to the historical outsourced personnel data; generating a first association fusion vector between multiple outsourced personnel associated nodes that have an association relationship with the outsourced personnel node according to the first target knowledge graph; constructing a second target knowledge graph according to the current outsourced personnel data; generating a second association fusion vector between multiple outsourced personnel associated nodes that have an association relationship with the outsourced personnel node according to the second target knowledge graph; replacing the first association fusion vector in the first target knowledge graph according to each of the second association fusion vectors to obtain a corresponding updated association vector; The step of replacing the first association fusion vector in the first target knowledge graph according to the second association fusion vector to obtain a corresponding updated association vector includes: Acquire multiple first entities and second entities in the first target knowledge graph and the second target knowledge graph, and acquire corresponding first and second relationships based on each of the first entities and the second entities; acquire corresponding entity similarity based on each of the first entities and the second entities; acquire corresponding relationship similarity based on each of the first relationships and the second relationships, and acquire corresponding entity relevance based on each of the relationship similarities and entity similarities; acquire total similarity based on the multiple entity relevances, and acquire the difference between the first associated fusion vector and the second associated fusion vector based on the total similarity; if the difference is greater than a preset threshold, mark the first associated fusion vector; if the difference is less than the preset threshold, replace the first associated fusion vector in the first target knowledge graph with the second associated fusion vector to obtain an updated associated vector; Determine the actual weight between each outsourced personnel node and multiple outsourced personnel associated nodes according to the association relationship, and add each of the actual weights to the corresponding updated association vector to obtain a corresponding weighted association vector; Training a preset matching prediction model according to the plurality of weighted association vectors to obtain a standard matching prediction model; A query request is obtained and input into a job matching prediction model to make an outsourced personnel recommendation.

2. The intelligent management method for outsourcing labor services based on user agreements according to claim 1 is characterized in that: Before the step of generating, according to the target knowledge graph, an updated association vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node, the step includes: Obtaining, from the target knowledge graph, M initial node groups corresponding to each outsourced personnel node, wherein the Nth initial node group includes N-1 feature nodes associated with the outsourced personnel node, where N is a positive integer and N∈[1,M]; Obtaining a specified time, and removing characteristic nodes that do not fall within the specified time from the M initial node groups corresponding to the outsourced personnel nodes, to obtain M updated node groups; A feature vector is obtained according to the feature node, and the feature vectors of all feature nodes in each update node group are fused according to the outsourced personnel node to obtain a fused vector.

3. The intelligent management method for outsourcing labor services based on user agreements according to claim 1 is characterized in that: The step of inputting the query request into the job matching prediction model to recommend outsourced personnel includes: Obtaining request information according to the query request, and extracting multiple demand elements in the request information, wherein the demand elements are mapped to outsourced personnel associated nodes in the target knowledge graph; Inputting each of the demand elements into a work matching prediction model to obtain an actual associated fusion vector corresponding to each of the demand elements; Obtaining a corresponding correlation matching degree according to each of the actual correlation fusion vectors and the updated correlation vector, and sorting the multiple correlation matching degrees to obtain a matching degree sorting table; Selecting the outsourced personnel association node corresponding to the first-ranked association matching degree from the matching degree ranking table as the recommended association node; The corresponding outsourcing personnel node is obtained according to the recommended associated node, and the user is recommended according to the outsourcing personnel node.

4. The intelligent management method for outsourcing labor services based on user agreements according to claim 3 is characterized in that: After the step of acquiring the corresponding outsourcing personnel node according to the recommended associated node and recommending the user according to the outsourcing personnel node, the method further includes: Obtaining the feedback result of the outsourced personnel node's recommendation to the user, and determining whether the feedback result meets the preset expected value; If the feedback result reaches the preset expected value, the recommendation result is determined to be qualified; If the feedback result does not meet the preset expected value, returning to the step of selecting the outsourced personnel associated node corresponding to the first associated matching degree from the matching degree ranking table; The outsourced personnel association node corresponding to the second-ranked association matching degree is selected and re-used as the recommended association node until the feedback result reaches the preset expected value.

5. An intelligent management system for outsourcing labor services based on user agreements, characterized in that: include: A construction module, configured to obtain outsourced personnel data and construct a target knowledge graph based on the outsourced personnel data, wherein the target knowledge graph includes a plurality of outsourced personnel nodes and a plurality of outsourced personnel associated nodes corresponding to each outsourced personnel node; An acquisition module, configured to acquire a corresponding association relationship according to each of the outsourced personnel nodes and the outsourced personnel association nodes; A generation module, configured to generate, based on the target knowledge graph, an updated association vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node, including: obtaining historical outsourced personnel data and current outsourced personnel data based on the outsourced personnel data, and constructing a first target knowledge graph based on the historical outsourced personnel data; generating, based on the first target knowledge graph, a first association fusion vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node; constructing a second target knowledge graph based on the current outsourced personnel data; generating, based on the second target knowledge graph, a second association fusion vector between a plurality of the outsourced personnel associated nodes that have an association relationship with the outsourced personnel node; replacing the first association fusion vector in the first target knowledge graph with each of the second association fusion vectors to obtain a corresponding updated association vector; The step of replacing the first association fusion vector in the first target knowledge graph according to the second association fusion vector to obtain a corresponding updated association vector includes: Acquire multiple first entities and second entities in the first target knowledge graph and the second target knowledge graph, and acquire corresponding first and second relationships based on each of the first entities and the second entities; acquire corresponding entity similarity based on each of the first entities and the second entities; acquire corresponding relationship similarity based on each of the first relationships and the second relationships, and acquire corresponding entity relevance based on each of the relationship similarities and entity similarities; acquire total similarity based on the multiple entity relevances, and acquire the difference between the first associated fusion vector and the second associated fusion vector based on the total similarity; if the difference is greater than a preset threshold, mark the first associated fusion vector; if the difference is less than the preset threshold, replace the first associated fusion vector in the first target knowledge graph with the second associated fusion vector to obtain an updated associated vector; a determination module, configured to determine an actual weight between each outsourced personnel node and a plurality of outsourced personnel associated nodes according to the association relationship, and add each of the actual weights to a corresponding updated association vector to obtain a corresponding weighted association vector; A training module, configured to train a preset matching prediction model based on a plurality of weighted association vectors to obtain a standard matching prediction model; The recommendation module is used to obtain a query request and input the query request into the job matching prediction model to recommend outsourced personnel.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images