Project team recommendation method and system and computer equipment

Abstract

The invention discloses a project team recommendation method and system and computer equipment, and belongs to the technical field of electric digital data processing, and the method comprises the following steps: obtaining a knowledge graph based on personnel information and business information; obtaining a first member list according to the knowledge graph and the project information; based on the large model and the similarity, obtaining scores of the recommendation personnel and the item information of the first member list; and generating a recommendation list according to the score. Based on the knowledge graph and the large model, the project information and the personnel information are tested and understood, the recommendation list and the recommendation personnel are generated, the multi-dimensional characteristics of the personnel information can be more comprehensively considered, the recommendation accuracy is improved, and the recommendation effect is improved. The method can be applied to the technical field of artificial intelligence such as machine learning and natural language processing.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic digital data processing, and in particular to a project team recommendation method, system and computer equipment. Background Art

[0002] With the rapid development of artificial intelligence technology, large language models (LLMs) have become an important force in promoting progress in the fields of natural language processing (NLP) and machine learning. These large models can understand and generate natural language through deep learning and massive data training, showing unprecedented language understanding and generation capabilities.

[0003] In the field of project management and team collaboration, accurate project team and team member recommendations are crucial to improving project success rates, optimizing resource allocation, and enhancing team collaboration efficiency.

[0004] Traditional project team recommendations often rely on rule engines or simple matching algorithms. These methods have limitations in dealing with complex and changing team requirements and are difficult to capture the dynamic interactions between team members and the nuances of project requirements, resulting in limited accuracy and applicability of recommendations. In addition, with the rise of remote work and global teams, cross-cultural and cross-time zone team formation has become more common, further increasing the complexity of team recommendations.

[0005] Therefore, designing a project team recommendation method based on big models has become an important research direction. By utilizing the powerful semantic understanding and generation capabilities of big models and combining them with advanced machine learning technology, we can achieve in-depth analysis of project requirements and team member capabilities, thereby providing more accurate and personalized team recommendations. Summary of the invention

[0006] In view of the above technical problems existing in the prior art, the present invention provides a project team recommendation method, system and computer device, which provide more effective team recommendation and formation solutions in a changing project environment.

[0007] The first aspect of the present invention discloses a method for recommending a project team, comprising the following steps: obtaining a knowledge graph based on personnel information and business information; obtaining a first member list based on the knowledge graph and project information; obtaining scores of recommended personnel and project information of the first member list based on a large model and similarity; and generating a recommendation list based on the scores.

[0008] Preferably, the method for constructing a knowledge graph includes: Collect business and personnel information; Extracting entities and relationships based on the business information and personnel information; Based on the entities and relationships, a business opportunity-personnel knowledge graph is generated.

[0009] Preferably, the entities include: business opportunity project entities and personnel entities; The attributes of the business opportunity project entity are selected from the following indicators: business type, strategic value, technical requirements, project establishment method, bidding method and construction and maintenance method; The attributes of the person entity are selected from the following indicators: skills, experience, position, soft skills and current project status; The relationships include: Suitable: a person is suitable for a certain position in a certain type of business; Participated: a person has participated in a certain business opportunity project; Demand: a certain business opportunity project requires a certain technology or skill.

[0010] Preferably, the business information includes structured information and long text information; The structured information, long text information and personnel information are vectorized and inserted into the knowledge base of the knowledge graph.

[0011] Preferably, the method for searching the first member list includes: Extract long text information and / or structured information from project information; According to the structured information, accurate matching search is performed in the knowledge graph / knowledge base to obtain the second member list; According to the long text information, a second vector representation is obtained; based on the similarity, related entities of the second vector are searched in the knowledge graph, and the related entities include business opportunity project entities and personnel entities; Combine the knowledge graph and related entities for logical reasoning to obtain the third member list; The second member list and the third member list are combined to obtain a first member list.

[0012] Preferably, the second member list and the third member list are combined to obtain the first member list.

[0013] Preferably, based on the project status of the personnel entity, the recommended personnel who do not meet the production process are excluded from the first member list; Excludes personnel whose project status is not available or who are responsible for more than the threshold number of projects.

[0014] Preferably, based on the big model, a recommendation reason is generated for the recommender on the recommendation list; or According to the evaluation result of the recommendation list, the strategy for obtaining the first member list is adjusted, or the large model is fine-tuned.

[0015] The second aspect of the present invention provides a recommendation system for a project team for implementing the above-mentioned recommendation method, including: a knowledge module, a search module and a recommendation module, wherein the knowledge module is used to obtain a knowledge graph based on personnel information and business information; the search module is used to obtain a first member list based on the knowledge graph and project information; the recommendation module is used to obtain scores of recommended personnel and project information of the first member list based on a large model and similarity; and generate a recommendation list based on the scores.

[0016] A third aspect of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a program; the processor is used to execute the program, and when the program is executed, the above-mentioned recommendation method is implemented.

[0017] Compared with the prior art, the beneficial effects of the present invention are: based on the knowledge graph and the big model, the project information and personnel information are tested and understood, and a recommendation list and recommended personnel are generated, which can more comprehensively consider the multi-dimensional characteristics of personnel information, improve the accuracy of the recommendation, and enhance the recommendation effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of the project team recommendation method of the present invention; Figure 2 is a flow chart of a method for obtaining a first member list; Figure 3 is a logic block diagram of the recommendation system of the present invention; Figure 4 It is a computer device framework diagram of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Overview: The technical solution provided by the present invention is mainly implemented by large-scale model technology. The large model here refers to a deep learning model with large-scale model parameters, which can usually contain hundreds of millions, tens of billions, hundreds of billions, trillions or even more than ten trillion model parameters. The large model can also be called a foundation model / foundation model. The large model is pre-trained with large-scale unlabeled corpus to produce a pre-trained model with more than 100 million parameters. This model can adapt to a wide range of downstream tasks, and the model has good generalization ability, such as large-scale language model (LLM), multi-modal pre-training model, etc.

[0021] Artificial Intelligence (AI): It is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines, so that machines have the functions of perception, reasoning and decision-making.

[0022] Natural language processing is an important direction in the fields of computer science and artificial intelligence. It studies various theories and methods that can achieve effective communication between people and computers using natural language. Natural language processing is a science that integrates linguistics, computer science, and mathematics. Therefore, research in this field will involve natural language, that is, the language people use in daily life, so it is closely related to the study of linguistics. Natural language processing technology usually includes text processing, semantic understanding, machine translation, robot question answering, knowledge graph and other technologies.

[0023] Knowledge Graph is a structured tool for representing and displaying knowledge. It consists of nodes (entities) and edges (relationships). Nodes represent entities, and edges represent relationships between entities. It uses visualization technology to present knowledge in a graphical way to help users better understand and use knowledge. Knowledge graph aims to describe concepts, entities, events and the relationships between them in the objective world. It is an advanced technical means that can be used to store and manage various knowledge and to intuitively display various knowledge.

[0024] The present invention is further described in detail below in conjunction with the accompanying drawings: like Figure 1As shown, the first aspect of the present invention provides a project team recommendation method based on a large model, comprising the following steps: Step 101: Build a knowledge graph based on personnel information and business information.

[0025] Personnel information and business information are collected from historical data.

[0026] Step 102: Obtain a first member list based on the knowledge graph and project information.

[0027] Project information is the input business information, including project requirements such as business opportunity description.

[0028] Step 103: Based on the large model, the similarity between the first member list and the project information and the recommendation score are calculated, and a recommendation list is generated, wherein the recommendation list includes at least one recommender.

[0029] Step 104: Based on the big model, generate recommendation reasons for the recommenders in the recommendation list.

[0030] Step 105: Obtain the test results based on the recommendation list and the reasons for recommendation.

[0031] Step 106: According to the test results, fine-tune the large model or obtain a strategy for the first member list.

[0032] Based on the knowledge graph and big model, the project information and personnel information are tested and understood, and the recommendation list and recommended personnel are generated. This can more comprehensively consider the multi-dimensional characteristics of personnel information, improve the accuracy of recommendations, and enhance the recommendation effect. It can handle more complex data patterns, understand more subtle language differences, and take into account more contextual information, so as to provide more effective team building suggestions in a changing work environment. It can comprehensively consider factors such as positions, basic information, personnel idleness, and production processes, so that the recommendation results are more in line with actual needs and improve the practicality of project team recommendations.

[0033] The large models in steps 103 and 104 can be the same large model or different types of large models, which can be selected according to actual test results.

[0034] In step 101, the knowledge graph is constructed by the following steps: Step 201: Historical data collection and organization.

[0035] Business data / project data: Extract business information from existing project documents (such as business opportunity project documents), databases, and related business systems. For example, obtain data on fields such as business type, strategic value score, and business opportunity description from the Excel table of the business opportunity project table.

[0036] Perform natural language processing (NLP) on the business opportunity description to extract key technical requirements and business features. For example, through lexical analysis, syntactic analysis, and named entity recognition, we can extract technical requirements such as "cloud computing", "5G", "Internet of Things", "big data" and business features such as "field manager system" from the description of "upgrading the field manager system based on customer needs using new generation communication technologies such as cloud computing, 5G, Internet of Things, and big data".

[0037] Personnel data: Integrate personnel information in the human resources management system, including basic skills, project experience (names of projects participated in, roles held, etc.), position information (current position, position level, etc.) and soft skills assessment (evaluation results of communication skills, teamwork ability, etc.).

[0038] Step 202: Define entities and relationships.

[0039] Entities: Identify the main entity types in the knowledge graph, including opportunity project entities and person entities.

[0040] For a business opportunity project entity, its attributes may include business type, strategic value, technical requirements (extracted from the business opportunity description), project establishment method, bidding method, construction and maintenance method, etc. For example, a business opportunity project entity may have attributes such as "ICT" business type, "high" strategic value, and "cybersecurity" technical requirements.

[0041] For a person entity, its attributes may include skills, experience, position, soft skills, current project status (whether available, number of projects in charge, etc.), etc. For example, a person entity may have the skill of "cybersecurity expert", the experience of "participated in 5 ICT projects", the position of "project manager", and the soft skill of "high communication ability".

[0042] For long texts such as business opportunity descriptions, the vector model can be used to convert them into vectors, combined with other structured information, and inserted into the knowledge base. For example, the vector insertion interface can be used to insert the business opportunity description vector into the knowledge base storage structure together with other business opportunity information. For personnel information, its multi-dimensional information is also organized, and some information may also need to be converted into vector form (such as skill vectors) before being inserted into the knowledge base.

[0043] Relationship: defines the relationship between entities, such as the "suitable" relationship, which indicates that a person is suitable for a certain position in a certain type of business; the "participated" relationship, which indicates that a person has participated in a certain business opportunity project; the "required" relationship, which indicates that a certain business opportunity project requires a certain technology or skill.

[0044] The constructed knowledge graph can also be called: business opportunity-personnel knowledge graph. When new information is updated, the graph structure and node attributes are updated in time.

[0045] Step 203: Knowledge graph storage and display.

[0046] You can use a graph database (such as Neo4j) to store knowledge graphs. Graph databases are specifically designed to process relational data of nodes and edges, and can perform graph query and traversal operations efficiently.

[0047] In a graph database, business opportunity project entities and personnel entities are nodes, and the relationship between them is an edge. For example, there is an edge of "suitable" relationship between a personnel node with cybersecurity skills and a business opportunity project node with cybersecurity requirements.

[0048] You can use graphical tools (such as Neo4j Browser) to intuitively view the structure of the knowledge graph. Nodes are represented by circles or squares to represent different types of entities, and edges are represented by lines to represent the relationship between entities, with the name of the relationship labeled. For example, in the graphical representation, you can see a relationship display like "Personnel A - Suitable - Business Opportunity Project B".

[0049] When new information is updated, the graph structure and node attributes of the knowledge graph can be updated in a timely manner.

[0050] like Figure 2 As shown, in step 102, the method for searching the first member list includes: Step 301: extracting long text information and / or structured information from project information.

[0051] Structured information includes clear fields such as business type and strategic value score. When the business opportunity description of new project information is input, it needs to be vectorized. After vectorization, the project information can be searched in the knowledge graph.

[0052] Step 302: Based on the structured information, perform precise matching search in the knowledge graph / knowledge base to obtain a second member list.

[0053] Step 303: Obtain a second vector representation based on the long text information; based on the similarity, search for related entities of the second vector in the knowledge graph.

[0054] Related entities include the Opportunity Project entity and the Person entity.

[0055] Step 304: Perform logical reasoning based on the knowledge graph and related entities to obtain a third member list.

[0056] The rules of logical reasoning can be preset based on experience, and before use, relevant rules are selected based on the similarity of business information.

[0057] Step 305: Combine the second member list and the third member list to obtain a first member list.

[0058] More specifically, when the project information contains both structured information and long text information, the same position in the second member list and the third member list can be combined; different positions can be combined as the first member list, but not limited to this.

[0059] Step 306: exclude personnel who do not meet the production process from the first member list according to the project status of the personnel entity. For example, if the project status of the personnel entity is not idle, or the number of projects in charge is greater than a threshold, the recommended personnel will be excluded.

[0060] The recommendation list includes positions and the corresponding recommended persons.

[0061] In step 103, a training set can be constructed to preset scores for the matching between the business information and members in the training set, and the training set can be used to pre-train or fine-tune the large model. By combining the large model with the knowledge graph, the matching and similarity scores between the project information and the recommended persons in the first member list can be scored. The recommended persons can be selected from the first member list by maximizing the scores.

[0062] In similarity calculation, entity similarity and rule similarity strategies are combined. For example, the similarity between personnel and business opportunities in terms of skills and rule requirements is calculated. For long text type business opportunity descriptions, the vector model + large model is used to calculate the recommendation score, comprehensively considering the matching degree between personnel skills, experience and business opportunity requirements.

[0063] In step 104, a prompt word template may be preset, combined with the prompt word template and the recommender, to generate a prompt word, and the prompt word may be input into the big model to obtain the recommendation reason fed back by the big model.

[0064] In a specific test, an interface docking solution was introduced. Backend interfaces, insertion interfaces, vector search interfaces, etc. were deployed. Other related systems such as the human resources management system (to obtain personnel status information) were docked through corresponding interfaces to ensure the real-time and accuracy of information.

[0065] In step 106, the process and related strategies are continuously optimized through test experiments. If the recommendation results are found to be inaccurate, it is analyzed whether it is a problem with the similarity / score calculation or the search strategy of the first member list, and targeted adjustments can be made. If it is due to the incompatibility of logic and scenarios, the interface design and logic can be adjusted in time to ensure that the entire process can meet the needs.

[0066] In one embodiment, a human assistant is designed to implement the above method. The human assistant receives requests and questions recommended by the project team through a dialogue interface. The request includes structured information and / or long text information. The human assistant also provides feedback to the user through the dialogue interface on the recommendation list and the reasons for the recommendation, thereby realizing human-computer interaction in a natural language manner.

[0067] Example 1. Industry: Education. Input: Purchase network equipment for Shandong University's information-based teaching and data center construction. The equipment is required to meet relevant technical specifications, such as load balancing equipment must have functions such as server performance optimization, and IT operation and maintenance management systems must have multiple monitoring and management functions. Recommend suitable project members and give corresponding reasons for the recommendation. Output: Based on the user's problem description, the following project members are recommended: Project Manager: LXY; Solution Manager: Huang; Bidding Manager: Wu; Account Manager: WYH. The reasons are as follows: Project Manager LXY served as the project manager in the 2021 Jinan Campus Campus Network Upgrade and Reconstruction Project of Shandong University and has rich experience in the education industry. Solution Manager Huang served as the solution manager in the 2021 Jinan Campus Campus Network Upgrade and Reconstruction Project of Shandong University and has a deep understanding of network equipment and technical specifications. Bidding Manager Wu served as the bidding manager in the 2021 Jinan Campus Campus Network Upgrade and Reconstruction Project of Shandong University and is familiar with the bidding process and customer needs. Account Manager WYH served as the account manager in the 2021 Jinan Campus Campus Network Upgrade and Reconstruction Project of Shandong University and has a good cooperative relationship with Shandong University. The above personnel all have relevant industry experience and technical capabilities and can meet user needs. Test results: reasonable.

[0068] Example 2. Industry: Government Affairs. Input: Procurement for the 5G + Rural Governance Service Procurement Project in Yangtun Town, Liaocheng City. Including agricultural Internet of Things base and environmental monitoring system, Ping An Smart Eye, hardware integration services, 5G wireless signal customization enhancement, various sensors, smart gateways, smart control cabinets and other hardware. Industry: Government Affairs. Output: According to the user's problem description, the following personnel are recommended: Project Manager: LA; Solution Manager: Liu; Bidding Manager: Zhou; Account Manager: WC; The reasons are as follows: Project Manager LA has participated in relevant technology application projects and served as a project manager in the 5G + Rural Governance Service Procurement Project in Yangtun Town, Liaocheng City, and has rich project management experience. Solution Manager Liu has relevant experience in agricultural Internet of Things base and environmental monitoring system, Ping An Smart Eye, hardware integration services, 5G wireless signal customization enhancement, various sensors, smart gateways, smart control cabinets and other hardware, and can provide professional technical support for the project. Bidding Manager Zhou has rich experience in project bidding and can strive for better cooperation conditions for the project. Account Manager WC has relevant experience in the government affairs industry and can better understand customer needs and provide high-quality customer service for the project. Test result: unreasonable.

[0069] Example 3. Industry: Government affairs. Input: Procurement for the 5G + rural governance service procurement project in Yangtun Town, Liaocheng City. Including agricultural IoT base and environmental monitoring system, Ping An Smart Eye, hardware integration services, customized enhancement of 5G wireless signals, various sensors, smart gateways, smart control cabinets and other hardware. Industry: Government affairs, recommend suitable project personnel and give reasons for the recommendation. Output: Based on the user's problem description, recommend the following personnel: Project Manager: XXR; Solution Manager: Hu; Bidding Manager: Lin Account manager: ZDL. The reasons are as follows: Project manager XXR has extensive experience in project management and can ensure the smooth progress of the project. Solution manager Hu has profound professional knowledge in smart platform construction and can provide strong technical support for the project. Bidding manager Lin has extensive experience in 5G business requirements, data transmission, positioning, etc. and can provide professional technical support for the project. Account manager ZDL has a deep industry background in the government affairs industry and can provide strong market support for the project. Test results: reasonable.

[0070] Example 4. Industry: Government affairs. Input: Information system integration implementation service project of Jinan Big Data Bureau, quantum communication application platform of e-government external network authentication system. Recommend suitable project personnel and give reasons for the recommendation. Output: Based on the user's problem description, the following personnel are recommended: Project Manager: YLJ; Solution Manager: Zhao; Bidding Manager: Gao; Account Manager: ZYJ. The reasons are as follows: Project Manager YLJ and Account Manager ZYJ have relevant experience in the industry (government affairs) involved in the project description, and have successful cases in the project type (ICT). Solution Manager Zhao and Bidding Manager Gao have relevant experience in the technical applications (cloud computing, big data, Internet of Things) involved in the project description, and have successful cases in the project type (ICT). In summary, the above recommended personnel have relevant experience in industry, project type and technical application, and can meet the project requirements. Test results: reasonable.

[0071] Example 5. Industry: Transportation. Input: Build a freight platform for the Dongchangfu District Transportation Bureau. Procurement details overview: Includes development, deployment and debugging services for the basic platform (applets, apps, web terminals, cloud service terminals) and business segment-related software. Output: Based on the user's problem description, the following personnel are recommended: Project Manager: WXS; Scheme Manager: Lin; Bidding Manager: Zhang; Account Manager: YWC.

[0072] Reason: The above personnel have rich experience in platform projects with ICT as the project type, and have a high degree of matching in the transportation industry. Among them, project manager WXS has served as a project manager in the construction project of the city operation management platform of the urban construction center and has relevant experience. Test result: reasonable.

[0073] Example 6. Industry: Education. Input: Carry out the "four one" security project construction for primary and secondary schools and kindergartens, covering the purchase and installation of security facilities and equipment such as monitoring and alarm, with schools and kindergartens as the implementation entities, self-raised funds, simplified procedures, and special matters handled in a special manner. Appoint Ma as the bidding manager to follow up. Output: Based on the user's problem description, recommend the following personnel: Project Manager: WJ; Program Manager: Guo Bidding manager: Ma; Account manager: ZJP. The reasons are as follows: Project manager WJ and solution manager Guo have experience in participating in the construction of smart campuses, including campus card and security monitoring, in the ICT project of Zhonghua Martial Arts School in Laizhou, Shandong Province, which matches the procurement and installation of security facilities and equipment in the user's problem description.

[0074] Bidding manager Ma is responsible for bidding in the ICT project of Zhonghua Wushu School in Laizhou, Shandong Province, and matches the designated bidding manager Ma in the user's problem description. Account manager ZJP is responsible for customer relationship management in the ICT project of Zhonghua Wushu School in Laizhou, Shandong Province, and matches the implementation subject of schools and kindergartens in the user's problem description. In summary, the above recommended personnel all have relevant industry experience and project role capabilities, which meet the requirements of the user's problem description. Test results: reasonable.

[0075] Example 7. Industry: Education. Input: Procurement of Campus Network Monitoring and Broadcasting System Package A for Wentian Middle School Education Equipment Procurement Project. Including campus monitoring system (cameras, switches, etc.), campus broadcasting system (control center, software, speakers, etc.), campus network system (firewall, network management software, etc.) and dormitory air conditioners, covering equipment procurement, installation and commissioning and related services. The project manager among the recommended members needs to understand the education industry. Output: The recommended personnel are as follows: Project Manager: ZRH; Project Manager: Liu; Bidding manager: Gao; Account manager: LB. Reason: The above personnel have relevant project experience in the education industry and have high scores in project type and role ability matching. Test result: reasonable.

[0076] Example 8. Industry: Government affairs. Input: The customer Jinan Hongtai Medical Management Group Co., Ltd. has newly purchased a 6-story office building and has requirements for integrated wiring, wireless AP, DingTalk access control, etc. Output: Based on the user's problem description, the following personnel are recommended: Project Manager: WCB; Solution Manager: Yang; Bidding Manager: Li Chao Account manager: ZCH. Reason: These four people have project experience involving integrated wiring, wireless AP, DingTalk access control and other needs, and they have high scores in industry matching and role ability matching. Test result: reasonable.

[0077] Example 9. Industry: Medical. Input: Shandong Provincial Medical Insurance Bureau medical insurance information sharing and exchange subsystem project budget of 342.42. Through the construction of the information sharing and exchange subsystem, all information systems in Shandong Provincial Medical Insurance Bureau will be managed in a coordinated manner, data and services will be completely separated, and one-to-many service calls and many-to-one service collaboration will be realized. Output: Recommended personnel are as follows: Project Manager: SZY; Program Manager: Zhu; Bidding manager: Zhou; Account manager: SCC. Reason: The above personnel have played corresponding roles in the medical insurance information sharing and exchange subsystem project of Shandong Provincial Medical Insurance Bureau, and have experience in related projects of the Medical Insurance Bureau, which meets the requirements of the user's problem description. Test results: reasonable.

[0078] Example 10. Industry: Medical. Input: Jinan Oncology Hospital National Regional Medical Center Construction Disaster Recovery Data Center. Overview of procurement details: Including dedicated cloud resources, disaster recovery systems, storage switches, business switches, cloud platform services and disaster recovery services, IDC computer rooms and links and other equipment, as well as related technical support, operation and maintenance services, etc., are required to comply with relevant technical standards and meet the hospital's data storage, backup, recovery and other needs. Output: Based on the user's problem description, the following personnel are recommended: Project Manager: SZY; Solution Manager: Huang; Bidding Manager: Zhao; Account Manager: SCC. The reasons are as follows: Project Manager SZY has relevant industry experience and has participated in similar projects; Solution Manager Huang has relevant technical application experience and is familiar with disaster recovery systems and cloud platform services; Bidding Manager Zhao has relevant industry experience and has participated in similar projects; Account Manager SCC has relevant industry experience and is familiar with the needs of the medical industry. In summary, the above personnel all have relevant industry and technical experience and can meet user needs. Test results: reasonable.

[0079] Example 11. Input: This is my project requirement. Please recommend suitable project participants: Solution industry: Education; Business type: Big data. Output: Project manager: AW; Solution manager: Huang; Bidding manager: Li. Input: Please recommend an account manager. Output: Account manager: ZGD. Test result: Reasonable.

[0080] Example 12. Input: You are a project leader and need the current business opportunity requirements. Based on the historical business opportunity information, recommend suitable project members, including: project manager, solution manager, bidding manager and account manager. The project requirements are as follows: "City": "Qingdao", "Solution Industry": "Education", "Business Type": "ICT", "Expected Contract Amount (Yuan)": 300,000, "Business Opportunity Description", "Professional Teaching Resource Construction and Application Platform System Procurement", "Strategic Value Score": 66, "Project Establishment Method": "Ordinary Project Establishment", "Bidding Method": "Invitation to Bid", "Construction and Maintenance Method": "Construction and Maintenance". Output: Project Manager: WD, Solution Manager: Zhao, Bidding Manager: Wu, Account Manager: WRR. Test Result: Reasonable.

[0081] After the above test evaluation, the rationality of the recommendation exceeds 90%, which meets the use requirements. It can receive requests based on structured information, requests based on long text information, or a combination of the two. The method of the present invention can give efficient and accurate recommendations for project team members, improving employment efficiency and operational benefits.

[0082] The present invention utilizes the powerful semantic understanding and generation capabilities of large models, combined with advanced machine learning technology, to achieve in-depth analysis of project requirements and team member capabilities, thereby providing more accurate and personalized team recommendations. This large model-based method can handle more complex data patterns, understand more subtle language differences, and take into account more contextual information, thereby providing more effective team building suggestions in a changing work environment. It brings innovative solutions to the fields of project management and team collaboration, improves the efficiency and effectiveness of team building, and ultimately promotes project success. It breaks the traditional personnel recommendation algorithm construction model, introduces real-time personnel data, improves the accuracy of model recommendation results, and improves the actual application effect of the model. The project personnel recommendation list can be pushed to the project leader in real time with one click, assisting in the rapid completion of the pre-sales project delivery team formation and improving the overall project delivery efficiency.

[0083] The vector search based on RAG in the present invention can quickly retrieve relevant information from the knowledge base and provide basic data for personnel recommendation. Combining strategies such as knowledge graph, entity similarity and rule similarity, the multi-dimensional characteristics of role information can be more comprehensively considered to improve the accuracy of recommendation. The vector model is combined with the big model, and the vector model is used to perform preliminary processing on long texts. The big model further understands and generates reasons for recommendations, and jointly improves the recommendation effect. Comprehensively considering factors such as positions, basic information, personnel idle status and production processes can make the recommendation results more in line with actual needs and improve the practicality of the system.

[0084] The second aspect of the present invention also provides a system for implementing the above-mentioned recommendation method, such as Figure 3As shown, it includes: knowledge module 1, search module 2 and recommendation module 3. The knowledge module 1 is used to obtain a knowledge graph based on personnel information and business information; The search module 2 is used to obtain a first member list according to the knowledge graph and project information; The recommendation module 3 is used to obtain the scores of the recommended personnel and project information of the first member list based on the large model and similarity; and generate a recommendation list according to the scores.

[0085] The recommendation module 3 is also used to generate recommendation reasons for the recommenders in the recommendation list based on the big model.

[0086] The recommendation system further comprises an optimization module 4, which is used to adjust the strategy for obtaining the first member list or to fine-tune the large model according to the evaluation result of the recommendation list.

[0087] A third aspect of the present invention provides a computer device 10, such as Figure 4 , including a processor 11 and a memory 12.

[0088] The processor 11 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 11 may be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), and programmable logic array (PLA). The processor 11 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a central processing unit (CPU); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor may be integrated with a graphics processing unit (GPU), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor may also include an artificial intelligence (AI) processor, which is used to process computing operations related to machine learning.

[0089] The memory 12 may include one or more computer-readable storage media, which may be non-transitory. The memory 12 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 12 is used to store a computer program, which is used to be executed by the processor 11 to implement the project team recommendation method provided by the method invention in this application.

[0090] Illustratively, the computer device 10 also includes other components, which can be understood by those skilled in the art. Figure 4 The structure shown in the figure does not constitute a limitation on the computer device 10, and the computer device 10 may include more or less components than shown in the figure, or combine some components, or adopt a different arrangement of components.

[0091] The present invention can be applied to artificial intelligence technology fields such as machine learning, natural language processing, and autonomous driving.

[0092] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A project team recommendation method, characterized in that: The following steps are involved: Obtain a knowledge graph based on personnel information and business information; Obtaining a first member list according to the knowledge graph and the project information, wherein the first member list includes at least one recommender; Based on the big model and similarity, obtain the scores of recommended personnel and project information; A recommendation list is generated based on the ratings.

2. The recommendation method according to claim 1, characterized in that: The construction methods of knowledge graph include: Collect business and personnel information; Extracting entities and relationships based on the business information and personnel information; A knowledge graph of business opportunities and personnel is generated based on the entities and relationships.

3. The recommendation method according to claim 2, characterized in that: The entities include: business opportunity project entity and personnel entity; The attributes of the business opportunity project entity are selected from the following indicators: business type, strategic value, technical requirements, project establishment method, bidding method and construction and maintenance method; The attributes of the person entity are selected from the following indicators: skills, experience, position, soft skills and current project status; The relationships include: Suitable for: a person is suitable for a certain position in a certain type of business; Participation: A person has participated in a business opportunity project; Demand: A business opportunity project requires certain technology or skills.

4. The recommendation method according to claim 3, characterized in that: The business information includes structured information and long text information; The structured information, long text information and personnel information are vectorized and inserted into the knowledge base of the knowledge graph.

5. The recommendation method according to claim 3, characterized in that: Methods for searching the first member list include: Extract long text information and / or structured information from project information; According to the structured information, accurate matching search is performed in the knowledge graph / knowledge base to obtain the second member list; According to the long text information, a second vector representation is obtained; based on the similarity, related entities of the second vector are searched in the knowledge graph, and the related entities include business opportunity project entities and personnel entities; Combine the knowledge graph and related entities for logical reasoning to obtain the third member list; The second member list and the third member list are combined to obtain a first member list.

6. The recommendation method according to claim 5, characterized in that: The second member list is combined with the third member list to obtain a first member list.

7. The recommendation method according to claim 5, characterized in that: According to the project status of the personnel entity, the recommended personnel who do not meet the production process are excluded from the first member list; Excludes personnel whose project status is not available or who are responsible for more than the threshold number of projects.

8. The recommendation method according to claim 1, characterized in that: Based on the big model, generate recommendation reasons for the recommenders in the recommendation list; According to the evaluation result of the recommendation list, the strategy for obtaining the first member list is adjusted, or the large model is fine-tuned.

9. A recommendation system for a project team, characterized in that: Used to implement the recommendation method according to any one of claims 1 to 8, the recommendation system comprises: a knowledge module, a search module and a recommendation module, The knowledge module is used to obtain a knowledge graph based on personnel information and business information; The search module is used to obtain a first member list according to the knowledge graph and project information; The recommendation module is used to obtain scores of recommended personnel and project information of the first member list based on the large model and similarity; and generate a recommendation list according to the scores.

10. A computer device, characterized in that: comprising a memory storing a program; When the program is executed, the recommendation method according to any one of claims 1 to 8 is implemented.

Images