Method and system for intelligently collecting and editing knowledge in work order system

Abstract

The invention discloses a method for intelligently collecting and editing knowledge in a work order system, and belongs to the technical field of computers. The method comprises the steps of obtaining a to-be-processed file; performing feature extraction on the to-be-processed file to obtain feature information; performing clustering analysis on the feature information to obtain a clustering result; extracting question and answer pairs from the to-be-processed file to obtain question and answer pairs; and obtaining a knowledge base according to the clustering result and the question-answer pairs. According to the method, large model information abstracts, cluster label generation and clustering methods are combined, and system error information, chat record information and the like are fused, so that intelligent information collection is realized. Compared with a traditional density clustering algorithm, the method has the advantage that the clustering analysis efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and system for intelligent collection and editing of knowledge in a work order system. Background Art

[0002] In modern enterprise operations, the work order system is an important tool for maintaining service quality and responding to customer needs. However, traditional work order processing methods often rely on manual operations, which are not only inefficient but also prone to errors. At the same time, it is difficult to automatically accumulate solutions / knowledge for work order solutions. Therefore, manual operation strategies are often adopted to identify hot issues / knowledge in work order analysis or keyword screening to achieve knowledge collection and editing, such as Figure 1 shown.

[0003] 1. The following is the difference between the knowledge collection and compilation method in the present invention and the "A method for intelligent extraction of small sample event elements based on clustering algorithm" (patent number: CN202410359544.0): The patent is a method for intelligent extraction of small sample event elements based on clustering algorithm, which uses the Word2Vec model to vectorize text data, clusters it through the K-means algorithm, and selects multiple samples with the highest information entropy similarity weight as representative samples of the cluster. It is mainly aimed at event extraction in small sample cases, and can also be used as a method for knowledge extraction. However, the patent directly processes text data, and there is a problem of low efficiency in clustering analysis.

[0004] 2. The knowledge collection and editing in "A method and system for generating question-answer pairs based on a large language model" (patent number: CN202410465219.2) is different from that in the present invention: the patent is a method for generating question-answer pairs based on a large language model, which mainly uses document knowledge data through triple enhancement and two-stage retrieval enhancement, and then uses a large model to generate question-answer pairs, which is a re-editing and processing of explicit knowledge. However, the patent directly processes explicit knowledge, and in order to take into account the influence of implicit knowledge, the generated question-answer pairs are not sufficiently relevant.

[0005] Existing knowledge collection and editing methods and systems have the following shortcomings in the hybrid search part:

[0006] (1) They often rely on preset rules and templates and cannot flexibly adapt to different work order contents and formats, and cannot recognize synonyms and similar expressions.

[0007] (2) When processing a large number of work orders, existing knowledge collection and editing methods and systems make manual analysis difficult and unable to quickly locate hot issues, resulting in low analysis efficiency.

[0008] (3) The analysis granularity is relatively coarse and not comprehensive enough to accurately quantify hot issues and knowledge, so the improvement brought to subsequent knowledge empowerment is poor.

[0009] (4) The analysis dimension is limited. It can only analyze through the work order content and the completed opinion, and cannot analyze and mine by combining information such as system error messages and chat records. Summary of the Invention

[0010] The purpose of the present invention is to provide a method and system for intelligent knowledge acquisition and editing in a work order system, which can efficiently analyze work order data of inventory / increment, integrate system error message information, chat record information, etc., and intelligently realize intelligent knowledge acquisition and editing, efficiently construct a work order problem system, and assist subsequent intention recognition and work order operation analysis. Through this method and system, an intelligent knowledge extraction pipeline is constructed to regularly complete automated knowledge extraction, reduce operation costs, improve the intelligent response effect, improve the iteration efficiency, provide a method and tool for intelligent knowledge acquisition and editing for the intelligent response system, and obtain an intelligent question and answer system that continuously evolves and improves.

[0011] To solve the above technical problems, the present invention provides a method for intelligent knowledge acquisition and editing in a work order system, including:

[0012] Obtain the file to be processed;

[0013] Extract features from the file to be processed to obtain feature information;

[0014] Perform cluster analysis on the feature information to obtain a clustering result;

[0015] Extract question and answer pairs from the file to be processed to obtain question and answer pairs;

[0016] Obtain a knowledge base according to the clustering result and the question and answer pairs.

[0017] Preferably, the file to be processed includes work order text, completed opinion, error message, and chat record.

[0018] Preferably, the feature information includes work order summary, work order type information, and conversation summary.

[0019] Preferably, extracting features from the file to be processed to obtain feature information specifically includes the following steps:

[0020] Extract a summary from the work order text to obtain a work order summary;

[0021] Identify the type of the work order text to obtain work order type information;

[0022] Extract a summary from the conversation summary to obtain a conversation summary.

[0023] Preferably, questions and answers pairs are extracted from the file to be processed to obtain the Q&A pairs, which specifically include the following steps:

[0024] Determine whether the work order type information of the characteristic information of the file to be processed is a consultation type;

[0025] If the work order type information is of the consultation type, extract questions and answers pairs from the work order text, completion opinion, error message, and chat record of the corresponding file to be processed to obtain the Q&A pairs.

[0026] Preferably, the following steps are further included:

[0027] Add labels to the clustering results.

[0028] The present invention also provides a system for intelligent knowledge acquisition and editing in a work order system, including:

[0029] An acquisition module, configured to acquire the file to be processed;

[0030] A large model text generation module, configured to perform feature extraction on the file to be processed to obtain feature information;

[0031] A clustering module, configured to perform clustering analysis on the feature information to obtain a clustering result;

[0032] A Q&A pair extraction module, configured to extract questions and answers pairs from the file to be processed to obtain the Q&A pairs;

[0033] A knowledge acquisition and editing module, configured to obtain a knowledge base according to the clustering result and the Q&A pairs.

[0034] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0035] The present invention creatively proposes a method for intelligent knowledge acquisition and editing in a work order system, making up for the deficiencies and pain points of the existing work order knowledge acquisition and operation analysis processes, improving the acquisition and analysis efficiency, and being able to intelligently and efficiently achieve knowledge precipitation. Verified based on the practical scenario of a certain project case, 400,000 work orders in the past year were analyzed and knowledge was acquired and edited. The effects in aspects such as consultation type work order identification, work order knowledge extraction, and intelligent work order analysis are as follows:

[0036] (1) Consultation type work order identification: Combining the self-generated instruction method, a high-quality dataset of nearly 5,000 pieces was constructed, and a special consultation type identification model was trained based on a 1.8B large model, with a precision rate of 78%, an increase of 46% in precision rate compared to the 72B large model.

[0037] (2) Work order knowledge extraction: Combine large models to intelligently construct standard questions and answers, achieve intelligent knowledge editing and compilation, improve the efficiency of FAQ knowledge editing and compilation, with an overall accuracy rate of 82%+, and the editing and compilation efficiency has been improved from 5 days / 100 articles to 0.5 days / 100 articles, a 10-fold increase in efficiency.

[0038] (3) Intelligent work order analysis method: Combine large model information summary, cluster label generation and clustering methods, integrate system error message, chat record information, etc., to achieve intelligent information collection. After model fine-tuning, the accuracy rate of work order summary generation reaches 92%, and the accuracy rate of cluster label generation reaches 82%. Compared with the traditional density clustering algorithm, the clustering analysis efficiency is increased by nearly 6 times. Brief Description of the Drawings

[0039] The following further elaborates on the specific implementation manners of the present invention in conjunction with the accompanying drawings.

[0040] Figure 1 is the traditional intelligent editing and processing flow chart;

[0041] Figure 2 is the implementation flow chart of the knowledge intelligent editing and compilation solution of the present invention;

[0042] Figure 3 is the clustering schematic diagram;

[0043] Figure 4 is the schematic diagram of the knowledge intelligent editing and compilation result. Specific Embodiments

[0044] Many specific details are set forth in the following description in order to provide a thorough understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific implementations disclosed below.

[0045] The terms used in one or more embodiments of this specification are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of this specification. The singular forms "a", "the" and "said" used in one or more embodiments of this specification and the appended claims are also intended to include the plural forms unless the context clearly dictates otherwise. It should also be understood that the term " / and / " used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more related listed items.

[0046] It should be understood that although the terms first, second, etc. may be used in one or more embodiments of this specification to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to determining".

[0047] The following further describes the present invention in detail with reference to the accompanying drawings:

[0048] The present invention provides a method for intelligent knowledge acquisition and editing in a work order system, including:

[0049] Obtain the file to be processed;

[0050] Extract features from the file to be processed to obtain feature information;

[0051] Perform clustering analysis on the feature information to obtain a clustering result;

[0052] Extract question and answer pairs from the file to be processed to obtain Q&A pairs;

[0053] Obtain a knowledge base according to the clustering result and the Q&A pairs.

[0054] Preferably, the file to be processed includes work order text, completion opinions, error messages, and chat records.

[0055] Preferably, the feature information includes work order summaries, work order type information, and conversation summaries.

[0056] Preferably, extracting features from the file to be processed to obtain feature information specifically includes the following steps:

[0057] Extract a summary from the work order text to obtain a work order summary;

[0058] Identify the type of the work order text to obtain work order type information;

[0059] Extract a summary from the conversation summary to obtain a conversation summary.

[0060] Preferably, extracting question and answer pairs from the file to be processed to obtain Q&A pairs specifically includes the following steps:

[0061] Judge whether the work order type information in the feature information of the file to be processed is of the consultation type;

[0062] If the work order type information is of the consultation type, then extract question and answer pairs from the work order text, completion opinions, error messages, and chat records of the corresponding file to be processed to obtain Q&A pairs.

[0063] Preferably, the following steps are further included:

[0064] Add labels to the clustering results.

[0065] The present invention also provides a system for intelligent knowledge acquisition and editing in a work order system, including:

[0066] An acquisition module, configured to acquire files to be processed;

[0067] A large model text generation module, configured to extract features from the files to be processed to obtain feature information;

[0068] A clustering module, configured to perform clustering analysis on the feature information to obtain clustering results;

[0069] A question-and-answer pair extraction module, configured to extract question-and-answer pairs from the files to be processed to obtain question-and-answer pairs;

[0070] A knowledge acquisition and editing module, configured to obtain a knowledge base according to the clustering results and the question-and-answer pairs.

[0071] The knowledge intelligent acquisition and editing method of the present invention creatively proposes an efficient and intelligent knowledge intelligent acquisition and editing method, which intelligently identifies consultation work orders through a large model and realizes intelligent extraction of question-and-answer pairs, making up for the deficiencies and pain points of the existing acquisition and editing processes.

[0072] The present invention adopts an intelligent work order analysis method, combines large model information summarization, cluster label generation and clustering methods, integrates system error information, chat record information, etc., realizes intelligent information collection, and intelligently constructs question labels to efficiently construct a work order problem system, improving the work order analysis efficiency.

[0073] The present invention improves the algorithm for text information clustering. It realizes small sample knowledge extraction through information summarization + vectorization + dbscan algorithm. At the same time, for cluster representative samples, it adopts large model technology for extraction and generates cluster category labels, improving the efficiency of clustering analysis.

[0074] The present invention realizes knowledge acquisition and editing from work order content, conclusion opinions, error information and chat records through strategies such as multi-group information summarization, information collection, consultation work order recognition, and question-and-answer pair extraction. It is a mining and analysis of implicit knowledge, and the generated question-and-answer pairs have greater business relevance.

[0075] To better illustrate the technical effects of the present invention, the following specific embodiments are provided to illustrate the above technical processes:

[0076] Example 1. A method for intelligent knowledge extraction and compilation in a work order system, which makes up for the deficiencies and pain points of the existing work order knowledge extraction and compilation process, constructs an intelligent knowledge extraction pipeline with a large model as the core technology, can flexibly apply new work order data and dialogue information data, automatically perform knowledge extraction, and achieve intelligent extraction and compilation; as Figure 2 shown.

[0077] The core processing modules include: large model text generation, clustering model, cluster label generation, consultation type determination, question and answer pair extraction, knowledge extraction and compilation, etc.

[0078] (1) Large model text generation module: Through the large model combined with prompt word design, analyze and extract the file to be processed to generate feature information, realize information compression, and improve the performance of the subsequent clustering module.

[0079] The files to be processed include work order text, completion opinions, error messages, chat records, etc., and the feature information includes relevant work order summaries, dialogue summaries, work order type information, etc.

[0080] (2) Clustering model: The clustering model module performs clustering analysis on the extracted information, and classifies similar work orders or dialogue information into one category. This helps to improve the efficiency and accuracy of subsequent knowledge extraction and compilation; as Figure 3 shown.

[0081] (3) Cluster label generation: The cluster label generation module generates corresponding labels for each clustering result for subsequent problem analysis and knowledge extraction and compilation.

[0082] (4) Consultation type determination: Determine whether the work order or dialogue information belongs to the consultation type. This helps to distinguish different types of work orders and improve the pertinence of knowledge extraction and compilation.

[0083] (5) Question and answer pair extraction: Extract question and answer pairs from consultation type work orders or dialogue information to form standard questions - answers as answer pairs, and achieve intelligent knowledge extraction.

[0084] (6) Knowledge extraction and compilation: Organize and edit the extracted question and answer pairs to form a standardized knowledge base; as Figure 4 shown.

[0085] It realizes the intelligent knowledge extraction and compilation in the work order system, and makes up for the deficiencies and pain points of the existing work order knowledge extraction and compilation process. This innovative method not only improves the efficiency and accuracy of knowledge extraction and compilation, but also creates an intelligent knowledge extraction pipeline for users, and improves the efficiency of subsequent knowledge operation.

[0086] In this embodiment, a comparative experiment was conducted. Taking 400,000 work orders and their completion opinions, chat records and other data as examples, three mainstream work order knowledge acquisition and editing schemes in the industry were compared respectively. Data of multiple topics (about 10,000 pieces of data for each topic) were selected and experimental analyses were carried out using the three schemes respectively. The advantages and disadvantages of this method were demonstrated through comparative analysis:

[0087] Table 1 Comparative Table of Scheme Effect Tests Through experimental evaluation, the comparative table of experimental effect evaluation in Table 1 was obtained. It can be analyzed that:

[0088] Through this embodiment, more dimensions of information can be intelligently analyzed, and the efficiency of overall knowledge extraction and clustering analysis can be improved. Compared with traditional manual methods and basic intelligent methods, the efficiency is improved by several times.

[0089] Through this embodiment, consultation problems in work orders can be intelligently identified with an accuracy rate of 78%. It can assist the acquisition and editing personnel to focus on analyzing consultation problems (accounting for about 20%), and can greatly improve the efficiency of knowledge acquisition and editing.

[0090] In several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules, modules or units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units, modules or components can be combined or integrated into another device, or some features can be ignored or not executed.

[0091] The unit may or may not be physically separated. The components shown as units may be a physical unit or multiple physical units, that is, they may be located in one place, or may be distributed to multiple different places. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0092] In addition, the functional units in each embodiment of the present invention can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0093] In particular, according to the embodiments disclosed in the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, an embodiment of the present disclosure includes a computer program product that includes a computer program carried on a computer-readable medium, and the computer program includes program code for performing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from a network through a communication section and / or installed from a removable medium. When the computer program is executed by a central processing unit (CPU), the above-mentioned functions defined in the method of the present invention are performed. It should be noted that the above-mentioned computer-readable medium of the present invention can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above.

[0094] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code that contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combinations of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0095] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for intelligent knowledge collection and editing in a work order system, characterized in that: include: Get the files to be processed; Perform feature extraction on the files to be processed to obtain feature information; Perform cluster analysis on the feature information to obtain clustering results; Extract question and answer pairs from the document to be processed to obtain question-answer pairs; Based on the clustering results and question-answer pairs, a knowledge base is obtained.

2. The method for intelligent knowledge collection and editing in a work order system according to claim 1, characterized in that: The files to be processed include work order text, completion opinions, error information and chat records.

3. The method for intelligent knowledge collection and editing in a work order system according to claim 2, characterized in that: The feature information includes a work order summary, work order type information, and a conversation summary.

4. The method for intelligent knowledge collection and editing in a work order system according to claim 3, characterized in that: Feature extraction is performed on the file to be processed to obtain feature information, which specifically includes the following steps: Extract the summary of the work order text to obtain the work order summary; Perform type identification on the work order text to obtain the work order type information; Perform summary extraction on the conversation summary to obtain the conversation summary.

5. The method for intelligent knowledge collection and editing in a work order system according to claim 4, characterized in that: Extracting question and answer pairs from the file to be processed to obtain question-answer pairs specifically includes the following steps: Determine whether the work order type information of the characteristic information of the file to be processed is a consulting type; If the work order type information is consulting, question and answer pairs are extracted from the work order text, completion opinions, error information and chat records of the corresponding pending files to obtain question and answer pairs.

6. The method for intelligent knowledge collection and editing in a work order system according to claim 5, characterized in that: The following steps are also included: Add labels to the clustering results.

7. A system for intelligent knowledge collection and editing in a work order system, used to implement the method for intelligent knowledge collection and editing in a work order system as claimed in any one of claims 1 to 6, characterized in that: include: An acquisition module is used to obtain files to be processed; The large model text generation module is used to extract features from the files to be processed and obtain feature information; Clustering module, used to perform clustering analysis on feature information and obtain clustering results; A question-answer pair extraction module is used to extract question and answer pairs from the files to be processed to obtain question-answer pairs; The knowledge collection and compilation module is used to obtain the knowledge base based on the clustering results and question-answer pairs.

Images