Enterprise comprehensive management method, system and equipment based on large model and storage medium

Abstract

The invention provides an enterprise comprehensive management method, system and device based on a large model and a storage medium, and belongs to the technical field of enterprise intelligent management, and the method comprises the steps: building a vertical domain knowledge base based on original enterprise-level comprehensive management regulation data; according to the constructed vertical domain knowledge base, establishing a knowledge retrieval model Agent; constructing an expert experience module MoE based on the question and answer pairs; and according to the established knowledge retrieval model Agent and the expert experience module MoE, the large model answers the user question and outputs an answer. According to the invention, the existing comprehensive management regulations in an enterprise and accumulated experience of experts can be effectively integrated, and rapid response and accurate solution of questions can be realized through an intelligent means.

Description

Technical Field

[0001] The present invention relates to the field of enterprise intelligent management technology, and in particular to a large model-based enterprise comprehensive management method, system, equipment and storage medium. Background Art

[0002] In modern corporate management, the general management department plays a crucial role and is an indispensable key department within the enterprise. Its responsibilities cover a wide range of aspects, including human resources management, comprehensive administrative affairs processing, resource procurement, and project management. The effective execution of these functions is crucial to maintaining the daily operations of the enterprise and promoting the realization of its strategic goals.

[0003] However, in actual operations, due to the wide range and complexity of the areas involved, comprehensive management experts often face the challenges of a large number of professional issues. These issues not only include routine business process consulting, but may also involve deeper strategic decision-making support.

[0004] Currently, when internal employees encounter such issues, they typically seek help directly from general management specialists. In this scenario, business experts need to provide answers or guidance based on the company's general management procedures and their expertise. For example, they need to answer questions like which department is responsible for R&D project approval and what the travel reimbursement process is. Answering these questions requires general management specialists to review their knowledge, integrate information, and even draw on past experience.

[0005] Despite their extensive experience and expertise, experts face a massive daily workload, which not only significantly consumes their time and energy but also can lead to incomplete or inaccurate answers due to the limitations of their individual experience. Furthermore, as businesses continue to expand and their business scope becomes increasingly diverse, the traditional approach of relying on manual effort to resolve all issues is no longer able to meet the demand for efficient and accurate service.

[0006] Existing solutions build intelligent question-and-answer assistants based on historical question-and-answer pairs, taking only expert experience into account. Specifically, they first collect past question-and-answer pairs from the comprehensive management department to build a database. Next, they extract keywords from the user's questions. Based on the correlation between the question-and-answer pairs and the keywords in the database, they retrieve highly similar questions. Finally, they present the answers to the questions in descending order of similarity. This solution is obtained by first matching keywords with the questions in the question-and-answer pairs to identify questions that are similar or identical to those in previous historical questions. The answers to these similar or identical questions are then obtained and presented as solutions to the user's questions.

[0007] However, these solutions rely heavily on keywords, and question-answering assistants cannot fully understand the user's semantics. Retrieval accuracy depends primarily on the frequency of keyword occurrence, and requires high accuracy for keywords in user questions. Changing keywords can easily lead to inaccurate retrieval and inability to obtain effective answers. Summary of the Invention

[0008] In response to the above problems, the present invention provides a large-scale model-based enterprise comprehensive management method, system, equipment and storage medium to assist comprehensive management personnel in solving enterprise management problems, improve the accuracy of obtaining answers to problems, and greatly improve the efficiency of comprehensive management work.

[0009] The present invention provides a comprehensive enterprise management method based on a large model, comprising: Build a vertical domain knowledge base based on original enterprise-level comprehensive management procedures data; According to the constructed vertical domain knowledge base, a knowledge retrieval model Agent is established; Based on question-answer pairs, construct expert experience module MoE; Based on the established knowledge retrieval model Agent and expert experience module MoE, the big model answers user questions and outputs answers.

[0010] As a further improvement of the present invention, the construction of a vertical domain knowledge base based on the original enterprise-level comprehensive management procedure data includes: Collecting the original enterprise-level integrated management procedure data, and performing heterogeneous data analysis on the collected original enterprise-level integrated management procedure data; Performing data enhancement on the parsed enterprise-level integrated management procedure data to obtain enterprise-level integrated management procedure enhanced data, and storing the enterprise-level integrated management procedure enhanced data; The enterprise-level integrated management procedure reinforcement data is vectorized to generate a vector matrix, and the vector matrix is stored in a vector database.

[0011] As a further improvement of the present invention, the collecting of the original enterprise-level comprehensive management procedure data and the heterogeneous data analysis of the collected original enterprise-level comprehensive management procedure data include using PaddleOCR to convert the table data in the original .doc format file together with the text data into .txt format.

[0012] As a further improvement of the present invention, the enterprise-level comprehensive management procedure data after analysis is enhanced to obtain enterprise-level comprehensive management procedure enhanced data, and storing the enterprise-level comprehensive management procedure enhanced data includes compressing overlong paragraphs and adding paragraph summaries.

[0013] As a further improvement of the present invention, the enterprise-level comprehensive management procedure data after analysis is enhanced to obtain enterprise-level comprehensive management procedure enhanced data. Storing the enterprise-level comprehensive management procedure enhanced data includes naming each file with keywords and using business domain isolation folders.

[0014] As a further improvement of the present invention, the vectorization of the enterprise-level integrated management procedure enhancement data, generating a vector matrix, and storing the vector matrix into a vector database includes slicing the vector data in the vector database after storing the vector matrix into the vector database.

[0015] As a further improvement of the present invention, according to the established knowledge retrieval model Agent and expert experience module MoE, the big model answers user questions and outputs answers, including using langchain to build a thinking chain CoT system based on the big model.

[0016] The present invention provides an enterprise comprehensive management system based on a large model, comprising a vertical domain knowledge base construction unit, a knowledge retrieval model Agent establishment unit, an expert experience module MoE construction unit, and a large model answering unit, wherein: The vertical domain knowledge base construction unit is used to construct a vertical domain knowledge base based on the original enterprise-level comprehensive management procedure data; The knowledge retrieval model Agent establishment unit is used to establish a knowledge retrieval model Agent according to the constructed vertical domain knowledge base; The expert experience module MoE construction unit is used to construct the expert experience module MoE based on the question and answer pairs; The large model answering unit is used to answer user questions based on the established knowledge retrieval model Agent and expert experience module MoE, and output answers.

[0017] The present invention provides an electronic device comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above-mentioned large model-based enterprise comprehensive management method.

[0018] The present invention provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the above-mentioned enterprise comprehensive management method based on a large model when executed by a processor.

[0019] The present invention provides an enterprise comprehensive management method, system, device and storage medium based on a large model, constructs a vertical domain knowledge base by adopting original enterprise-level comprehensive management procedure data; establishes a knowledge retrieval model Agent based on the constructed vertical domain knowledge base; constructs an expert experience module MoE based on question and answer pairs; based on the established knowledge retrieval model Agent and expert experience module MoE, the large model answers user questions and outputs answers, effectively integrating the existing comprehensive management procedures within the enterprise and the accumulated experience of experts, and realizing rapid response and accurate answers to problems through intelligent means; through natural language interaction, comprehensive management personnel or enterprise employees can obtain relevant information and help more quickly and accurately, further improving work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a flowchart of the enterprise comprehensive management method based on a large model according to an embodiment of the present invention.

[0021] Figure 2 It is a detailed flowchart of the question-and-answer implementation of the enterprise comprehensive management method based on a large model according to an embodiment of the present invention.

[0022] Figure 3 It is a structural diagram of a large-model-based enterprise integrated management system according to an embodiment of the present invention.

[0023] Figure numerals: 10, vertical domain knowledge base construction unit; 20, knowledge retrieval model Agent establishment unit; 30, expert experience module MoE construction unit; 40, large model answer unit. DETAILED DESCRIPTION

[0024] The following is a combination of specific embodiments and appendix Figure 1-3 The invention is described in detail so that those skilled in the art can more fully understand the purpose, features and effects of the invention.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as commonly understood by those skilled in the art to which the invention belongs. In the event that the definition of a term in the present invention conflicts with the meaning commonly understood by those skilled in the art to which the invention belongs, the definition in the present invention shall prevail.

[0026] In large enterprises, experts in the general management department often need to answer questions from the employee and department levels regarding administration, human resources, project management and other fields. Their work involves a lot of demand understanding, knowledge retrieval, literature reading, and summarization and induction, which often leads to questions not being answered in a timely manner or the answers being incomplete or wrong, affecting the efficiency of enterprise management.

[0027] The present invention provides a large-scale model-based enterprise comprehensive management method, system, equipment and storage medium, thereby improving the accuracy of problem solving.

[0028] Example 1 As a specific embodiment of the present invention, this embodiment provides an enterprise comprehensive management method based on a large model, referring to Figure 1 , such as 2, the specific steps are as follows: S100. Build a vertical domain knowledge base based on original enterprise-level comprehensive management procedures data; S200: Establish a knowledge retrieval model Agent based on the constructed vertical domain knowledge base; S300: Build an expert experience module MoE based on question-answer pairs; S400: Based on the established knowledge retrieval model Agent and expert experience module MoE, the large model answers the user's question and outputs the answer.

[0029] By building a vertical domain knowledge base, the original enterprise-level comprehensive management procedure data within the enterprise is converted into structured, easy-to-retrieve knowledge information. Subsequently, a knowledge retrieval model Agent is established for the vertical domain knowledge base, which can improve the efficiency and accuracy of knowledge retrieval and enable users to quickly obtain the required knowledge. Constructing an expert experience module MoE based on question-answer pairs can effectively utilize historical experience, thereby improving the quality of knowledge retrieval.

[0030] The large-scale model-based integrated enterprise management method of this invention effectively integrates existing internal integrated management procedures and the accumulated experience of experts, enabling rapid response and accurate answers to questions through intelligent means. Through natural language interaction, integrated management personnel or employees can more quickly and accurately obtain relevant information and assistance, reducing decision-making difficulty and further improving work efficiency.

[0031] S100 builds a vertical domain knowledge base based on original enterprise-level comprehensive management procedures data, including: S110: Collect the original enterprise-level integrated management procedure data and perform heterogeneous data analysis on the collected original enterprise-level integrated management procedure data. Specifically, in this embodiment, the collected original enterprise-level comprehensive management procedure data needs to be converted into data in a set format.

[0032] In this embodiment, the collected original enterprise-level comprehensive management procedures data is a .doc format file containing table data and text data. The original enterprise-level comprehensive management procedures data can be various company rules and regulations and manuals, such as an employee handbook for human resources and a project approval standard process manual for project management.

[0033] In other feasible embodiments, the collected original enterprise-level integrated management procedure data may also be in other formats, such as .pdf format files.

[0034] Afterwards, the original enterprise-level comprehensive management procedure data in the collected .doc format file is converted to obtain the enterprise-level comprehensive management procedure data in the set format. In this embodiment, the table data in the original .doc format file is converted into .txt format together with the text data. Preferably, PaddleOCR (an OCR tool library based on Paddle) is used to convert the table data in the original .doc format file into .txt format together with the text data.

[0035] Converting .doc format to .txt format can remove redundant formats in enterprise-level comprehensive management procedures data, reduce the complexity of data processing, and facilitate adjustment and optimization.

[0036] By converting the original enterprise-level comprehensive management procedures data in .doc format to .txt format and using PaddleOCR technology to process the tables and text data, we can not only simplify the data processing process but also improve the accuracy and efficiency of data extraction. PaddleOCR is based on a deep learning algorithm and has a high character recognition accuracy rate, which can better handle complex typesetting and handwriting.

[0037] Since .txt files usually take up less storage space, converting the table data in the original .doc format file together with the text data into .txt format can also help save storage resources and help better manage and utilize the internal original comprehensive management procedure data.

[0038] In other feasible embodiments, other conversion tools may also be used to convert the table data in the original .doc format file together with the text data into the .txt format.

[0039] Furthermore, the enterprise-level integrated management procedure data converted into .txt format is processed including: Remove special symbols, spaces, abnormal indentation, title indentation and other junk information; Modify error messages such as incorrect section numbers and typos.

[0040] Through the above processing, the standardization of enterprise-level comprehensive management procedure data in .txt format files can be improved, and the data quality can be higher.

[0041] S120: Perform data enhancement on the analyzed enterprise-level integrated management procedure data to obtain enterprise-level integrated management procedure enhanced data, and store the enterprise-level integrated management procedure enhanced data. The data enhancement of the parsed enterprise-level comprehensive management procedures data includes compressing overly long paragraphs and adding paragraph summaries. Overly long paragraphs often contain a large amount of information, making it easy to lose focus during the search process. By compressing and extracting key information, the paragraphs can be made more compact and the content more readable.

[0042] By compressing overlong paragraphs and adding paragraph summaries, the core content of the paragraph can be summarized, which can enhance the accuracy of subsequent retrieval.

[0043] Preferably, a large model is used to compress overlong paragraphs and summary paragraph content. The large model summarizes the paragraph content and fills it below the paragraph title. Using a large model to compress overlong paragraphs and summary paragraph content improves the efficiency of data enrichment.

[0044] Furthermore, the obtained enterprise-level comprehensive management procedure reinforcement data is stored, including: Name each file using a keyword; Use business domains to isolate folders, such as human resources, project management, procurement, and other business domains.

[0045] Using subject terms to name individual files improves the accuracy of large models when searching for references based on questions. Using business domain isolation folders ensures that data from each business domain is isolated at the source to avoid data mixing.

[0046] By using PaddleOCR and a large model for text processing in the preprocessing stage, retrieval accuracy is enhanced.

[0047] S130, vectorizing the enterprise-level integrated management procedure reinforcement data to generate a vector matrix, and storing the vector matrix in a vector database Specifically, in this embodiment, after obtaining the enterprise-level integrated management procedure reinforcement data, it is segmented, cleaned, and vectorized to generate a vector matrix, which is then stored in a vector database. The vector database can use the lightweight Chroma database, a database system specifically designed for processing and storing large-scale vector data. The Chroma database can be seamlessly integrated with machine learning and AI applications, supporting multiple vector indexing technologies. Even when processing millions or even billions of vector data, it can provide sub-second query response times, significantly improving the query efficiency of large-scale vector data.

[0048] In other embodiments, the vector database may also adopt the Milvus database. In this regard, this embodiment does not impose any specific restrictions on the vector database.

[0049] Furthermore, the enterprise-level comprehensive management procedure enhancement data is segmented, cleaned, and vectorized based on a natural language embedding model to generate a vector matrix that can be directly stored in a vector database. The natural language embedding model is the bge model (BAAI General Embedding), which is an open source general vector model specifically designed to provide support for various types of information retrieval and large language model retrieval enhancement applications. In other feasible embodiments, other types of natural language embedding models can also be used to segment, clean, and vectorize the enterprise-level comprehensive management procedure enhancement data to generate a vector matrix that can be directly stored in a vector database.

[0050] After storing the vector matrix in the vector database, the vector data in the database is sliced. The batch size is set to limit the maximum length of a single text, and the overlap size is set to limit the length of repeated context segments. Using a reasonable slicing strategy can reduce retrieval difficulty while ensuring context relevance, improving the consistency of retrieval results and the understanding quality of large models.

[0051] Furthermore, a vector database is generated for each business domain knowledge, and a single complete file is used as the minimum storage unit in each vector database for storage. The storage mode is continuous storage.

[0052] By creating an independent vector database for each business domain, we can effectively isolate data. This helps prevent data confusion between different business domains and also facilitates domain-specific searches during subsequent Q&A. Query requests can be directly directed to the database of the relevant business domain, avoiding the additional overhead of cross-domain queries and improving query efficiency.

[0053] Furthermore, the original enterprise-level comprehensive management procedure data is collected and updated regularly, and stored in the vector database after preprocessing and vectorization, so that the data in the vector database can reflect the latest management procedures of the enterprise, thereby ensuring that the knowledge in the constructed vertical domain knowledge base is updated regularly to improve the accuracy of the retrieval results and reduce misleading information caused by outdated data.

[0054] Furthermore, a search strategy for the vector database is determined. In this embodiment, the search strategy for the vector database is based on the similarity between the user question and the knowledge. Preferably, the similarity between the user question and the knowledge is calculated using Euclidean distance.

[0055] Furthermore, the retrieval strategy of the vector database is based on the similarity between user questions and knowledge, and is combined with a keyword-based retrieval strategy to construct a hybrid retrieval mode to ensure retrieval accuracy.

[0056] In this embodiment, the obtained enterprise-level comprehensive management procedure enhanced data vector database uses special data structures and algorithms to store and process vector data, making storage and retrieval operations very efficient; in addition, the vector database can perform accurate matching based on the similarity between vectors, which is very important for knowledge question-answering application scenarios.

[0057] S100 collects original enterprise-level comprehensive management procedure data, performs data preprocessing and data enhancement, and performs word segmentation, embedding and other operations on the processed data to build a vector database, thereby obtaining a vertical domain knowledge base.

[0058] S200 establishes a knowledge retrieval model Agent based on the constructed vertical domain knowledge base, including: Using Langchain technology, the vertical domain knowledge base is mounted, and a knowledge retrieval model agent capable of natural language interaction with integrated management personnel is constructed. This knowledge retrieval agent performs knowledge retrieval and returns reference information. Preferably, knowledge retrieval and reference functionality are developed using Langchain's Chains component. Using Langchain technology allows for rapid model construction, accelerating model building efficiency.

[0059] Specifically, the knowledge retrieval model agent can be used alone. In the knowledge retrieval model agent, by asking specific comprehensive management questions, direct knowledge results or guidance opinions are returned.

[0060] The prompt words are as follows: Accurately find information and answer questions in the comprehensive management knowledge base.

[0061] Always end your answers with "Thanks for your question! Keep asking!" Refer to the following: {context} Question: {question}.

[0062] Furthermore, after a user enters a question, reference information is returned. Specifically, by traversing all knowledge base documents and extracting the fragments with the highest semantic relevance to the question, the model returns them. The file name and paragraph information of the fragment are also returned, allowing users to directly view the reference fragment in the original reference. This improves the interpretability of the content generated by the model and provides users with a more intuitive and convenient way to conduct fine-grained search results.

[0063] Through the knowledge retrieval model Agent, knowledge retrieval can be realized and reference information can be obtained, which improves the interactive experience.

[0064] S300 builds the expert experience module MoE based on question-answer pairs, including: Based on expert experience, questions and answers are collected, and a database of question-answer pairs is constructed and stored in a structured manner. The database is a file system.

[0065] In the real production environment of enterprises, business experts often deal with occasional, low-frequency, and complex abnormal problems. The solutions and ideas for such problems are very valuable and require model reference and expert experience to learn from. Therefore, by collecting these questions and answers, more comprehensive question-and-answer support can be provided for large models.

[0066] Furthermore, the expert experience module MoE adopts the NLP information retrieval strategy, through word segmentation, pattern matching, and weighted averaging of the question content, and performs relevance matching based on the knowledge of the expert experience file system, and returns the retrieved highly relevant expert experience to the big model.

[0067] The MoE module matches expert experience through information retrieval strategy for reference by the large model. Preferably, the information retrieval of the MoE module is performed based on keywords using the TD-IDF algorithm.

[0068] This embodiment uses langchain technology, combined with a vertical domain knowledge base, to create a large model Agent that can interact with managers in natural language, and constructs an expert module MoE to learn historical question-and-answer experience to provide users with comprehensive question-and-answer data support.

[0069] S400 uses the established knowledge retrieval model Agent and expert experience module MoE. The big model answers user questions and outputs answers. This includes using langchain to build a thinking chain CoT system based on the big model to achieve intelligent scheduling of Agent and MoE: Construct a task routing CoT to identify user intent. Specifically, identify the business domain where the user's question is located, improve retrieval speed, reduce errors caused by global retrieval, and assign the user's question to the corresponding knowledge retrieval model agent to perform knowledge retrieval in the business domain where the user's question is located.

[0070] In a feasible embodiment, the prompt words are as follows: Your task is to identify the intent. Based on the content of {order}, understand its meaning and select the most similar intent from the list. The intents include: 1. Questions about <human resources>; 2. Questions about <project management>; 3. Questions about <general administration>; 4. Questions about <resource procurement>. Answer only 1, 2, or 3. Example 1: Recognize the sentence "How to establish an R&D project?" and answer: 2. Example 2: Recognize the sentence "What should I do if I need to purchase a batch of monitors?" and answer: 4. Your answer: .

[0071] Through task routing CoT, user questions can be accurately mapped to the business domain where the questions are located, and searches can be performed in the knowledge base of the corresponding business domain to avoid invalid searches.

[0072] Furthermore, it also includes building a knowledge integration CoT to integrate the knowledge in the vertical domain knowledge base and the expert experience in the file system, so that the large model can return more comprehensive answers.

[0073] In a feasible embodiment, the prompt words are as follows: Your question is about knowledge integration. Summarize the information in {order1} and {order2} and answer {context}. Your answer:

[0074] The CoT system includes task routing and knowledge integration to automatically identify user intent and integrate relevant knowledge. By building a CoT system, intelligent scheduling of agents and MoEs is achieved.

[0075] Based on the established knowledge retrieval model Agent and expert experience module (MoE), the S400 large-scale model employs hybrid retrieval technology to answer user questions, combining keyword matching with semantic understanding. This increases the likelihood of retrieving relevant documents, improving recall and precision. Furthermore, the large-scale model merges the reference content provided by the knowledge retrieval model Agent and the expert experience module (MoE) to provide an integrated solution, enhancing its comprehensiveness.

[0076] The present invention's large-scale model-based integrated enterprise management method leverages the semantic understanding capabilities of large-scale models to achieve large-scale, dense semantic searches within seconds, as well as hybrid searches based on semantics and expert knowledge. Formally, the integrated management system implements document retrieval, empirical learning, and knowledge integration based on large-scale models, directly improving the work efficiency of integrated management personnel.

[0077] Example 2 As a specific embodiment of the present invention, this embodiment provides an enterprise comprehensive management system based on a large model, referring to Figure 3 , including vertical domain knowledge base construction unit 10, knowledge retrieval model Agent establishment unit 20, expert experience module MoE construction unit 30, large model answer unit 40, wherein, The vertical domain knowledge base construction unit 10 is used to construct a vertical domain knowledge base based on the original enterprise-level comprehensive management procedure data; The knowledge retrieval model Agent establishment unit 20 is used to establish a knowledge retrieval model Agent according to the constructed vertical domain knowledge base; The expert experience module MoE construction unit 30 is used to construct an expert experience module MoE based on the question-answer pair; The large model answering unit 40 is used to answer user questions based on the established knowledge retrieval model Agent and expert experience module MoE, and output answers.

[0078] The enterprise comprehensive management system based on the big model of the present invention improves the management efficiency and service quality of the enterprise by integrating various enterprise-level management procedure data and expert experience within the enterprise through four core units: the vertical domain knowledge base construction unit 10, the knowledge retrieval model Agent establishment unit 20, the expert experience module MoE construction unit 30, and the big model answering unit 40. The vertical domain knowledge base construction unit 10 constructs the vertical domain knowledge base after processing and transforming the original enterprise-level comprehensive management procedure data, laying the foundation for subsequent knowledge retrieval and problem answering; the knowledge retrieval model Agent establishment unit 20 establishes a model Agent specifically for retrieving relevant knowledge information based on the vertical domain knowledge base, making information retrieval more accurate and efficient; the expert experience module MoE construction unit 30 uses the form of question-answer pairs to convert the expert experience accumulated within the enterprise into operational knowledge data, supplements the amount of searchable information, makes up for the lack of knowledge in the vertical domain knowledge base, and enhances the intelligence and practicality of the system; the big model answering unit 40 combines the resources provided by the first two units, utilizes the powerful processing capabilities of the big model, answers questions raised by enterprise users, and outputs answers.

[0079] The enterprise comprehensive management system based on large models of the present invention digitizes the complex management procedure data and expert experience within the enterprise, can quickly provide decision support for managers, greatly shortens the decision-making cycle, can significantly reduce the information processing burden of comprehensive management personnel, and improve work efficiency and decision-making quality; through the effective organization and utilization of management procedure data and expert experience, it can provide users with a more comprehensive decision-making method.

[0080] Example 3 As a specific embodiment of the present invention, this embodiment provides an electronic device, including a memory, a processor and a computer program stored on the memory, and the processor executes the computer program to implement the steps of the large model-based enterprise comprehensive management method of the above-mentioned embodiment one.

[0081] Example 4 As a specific embodiment of the present invention, this embodiment provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the enterprise comprehensive management method based on the large model in the above-mentioned embodiment 1 are implemented.

[0082] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the enterprise comprehensive management method based on a large model described in Example 1.

[0083] The enterprise comprehensive management method, system, equipment and storage medium based on the big model provided by the present invention can assist comprehensive management personnel in solving enterprise management problems, share a large amount of retrieval, reading and sorting work, thereby greatly improving the efficiency of comprehensive management work.

[0084] The above description is merely a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any modification or equivalent variation based on the technical essence of the present invention shall still fall within the scope of protection claimed by the present invention.

Claims

1. A comprehensive enterprise management method based on a large model, characterized in that: The method comprises: Build a vertical domain knowledge base based on original enterprise-level comprehensive management procedures data; According to the constructed vertical domain knowledge base, a knowledge retrieval model Agent is established; Based on question-answer pairs, construct expert experience module MoE; Based on the established knowledge retrieval model Agent and expert experience module MoE, the big model answers user questions and outputs answers.

2. The enterprise comprehensive management method based on a large model according to claim 1 is characterized in that: The construction of a vertical domain knowledge base based on original enterprise-level comprehensive management procedure data includes: Collecting the original enterprise-level integrated management procedure data, and performing heterogeneous data analysis on the collected original enterprise-level integrated management procedure data; Performing data enhancement on the parsed enterprise-level integrated management procedure data to obtain enterprise-level integrated management procedure enhanced data, and storing the enterprise-level integrated management procedure enhanced data; The enterprise-level integrated management procedure reinforcement data is vectorized to generate a vector matrix, and the vector matrix is stored in a vector database.

3. The enterprise comprehensive management method based on a large model according to claim 2 is characterized in that: The collecting of the original enterprise-level comprehensive management procedure data and performing heterogeneous data analysis on the collected original enterprise-level comprehensive management procedure data includes using PaddleOCR to convert the table data in the original .doc format file together with the text data into .txt format.

4. The enterprise comprehensive management method based on a large model according to claim 2 is characterized in that: The enterprise-level comprehensive management procedure data after analysis is enhanced to obtain enterprise-level comprehensive management procedure enhanced data, and storing the enterprise-level comprehensive management procedure enhanced data includes compressing overlong paragraphs and adding paragraph summaries.

5. The enterprise comprehensive management method based on a large model according to claim 4 is characterized in that: The enterprise-level comprehensive management procedure data after analysis is enhanced to obtain enterprise-level comprehensive management procedure enhanced data. Storing the enterprise-level comprehensive management procedure enhanced data includes naming each file with a keyword and using business domain isolation folders.

6. The enterprise comprehensive management method based on a large model according to claim 2 is characterized in that: The vectorizing of the enterprise-level integrated management procedure reinforcement data to generate a vector matrix and storing the vector matrix in a vector database includes slicing the vector data in the vector database after storing the vector matrix in the vector database.

7. The enterprise comprehensive management method based on a large model according to claim 1 is characterized in that: According to the established knowledge retrieval model Agent and expert experience module MoE, the big model answers user questions and outputs answers, including using langchain to build a thinking chain CoT system based on the big model.

8. An enterprise integrated management system based on a large model, characterized in that: It includes vertical domain knowledge base construction unit, knowledge retrieval model Agent establishment unit, expert experience module MoE construction unit, and large model answer unit, among which, The vertical domain knowledge base construction unit is used to construct a vertical domain knowledge base based on the original enterprise-level comprehensive management procedure data; The knowledge retrieval model Agent establishment unit is used to establish a knowledge retrieval model Agent according to the constructed vertical domain knowledge base; The expert experience module MoE construction unit is used to construct the expert experience module MoE based on the question and answer pairs; The large model answering unit is used to answer user questions based on the established knowledge retrieval model Agent and expert experience module MoE, and output answers.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory, wherein: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

10. 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 7 are implemented.

Images