A method and system for mining PQL based on large language model analysis process

Abstract

The application discloses a method and system for mining PQL based on a large language model analysis process, and relates to the technical field of process query language, and the method comprises the following steps: S1, establishing a large language model; S2, knowledge base and explanation preparation; S3, PQL analysis prompt; S4, PQL statement explanation; S5, local highlight display; and S6, PQL statement fine tuning. The method and system can parse the PQL statement input by the user by using the large language model, and convert the PQL statement into natural language text that is easy to understand, so as to help the user quickly understand the specific meaning of the PQL, thereby reducing the understanding requirement of the user for the PQL statement while facilitating the reading and understanding of the user by feeding back the natural language analysis result with high readability without manual intervention.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of process query language, in particular to a method and system for analyzing process mining PQL based on a large language model. BACKGROUND

[0002] PQL, which stands for Process Query Language, is a language specifically developed for business processes. It can be directly translated into machine-executable code, and is widely used in business scenarios such as optimizing execution efficiency, querying irregular processes, separating rights and responsibilities, matching three single, and work order allocation. Compared with traditional BI methods, it is not only easier but also more efficient.

[0003] Currently, it is difficult for general users to analyze and understand the specific meaning of PQL in process mining. Even experienced users need to spend a lot of time and effort to analyze and understand more complex PQL. In addition, the training cost related to PQL is high and the cycle is long.

[0004] Therefore, in view of the above, the existing structure and deficiencies are studied and improved, and a method and system for analyzing process mining PQL based on a large language model are proposed. SUMMARY

[0005] To overcome the shortcomings of the prior art, the present application provides a method and system for analyzing process mining PQL based on a large language model, which solves the problems raised in the background art.

[0006] To achieve the above purpose, the present application realizes the following technical scheme: a method for analyzing process mining PQL based on a large language model, which includes the following operations:

[0007] S1, establish a large language model:

[0008] Prepare the pre-set PQL statement and its corresponding explanation as an analysis sample to input the large language model for sample learning and training, so as to confirm the PQL statement structure relationship and output format;

[0009] S2, knowledge base and explanation preparation:

[0010] Prepare the relevant structure and explanation of each element in PQL as a knowledge base for analysis, to provide the specific meaning of each element type parameter in PQL statement and the information of corresponding field;

[0011] S3, PQL analysis prompt:

[0012] Combine the sample explanation and PQL analysis prompt template to generate PQL analysis prompts, and input the PQL statement provided by the user together with the PQL analysis prompts into the large language model;

[0013] S4, PQL statement interpretation:

[0014] The user selects a PQL statement and inputs it into the large language model. The large language model analyzes the PQL statement and determines the PQL statement structure relationship based on the analysis results of the PQL statement. The large language model searches the knowledge base based on the PQL statement structure relationship and returns the natural language interpretation of the PQL statement and the corresponding relationship label between the elements and the interpretation.

[0015] S5, partial highlight display:

[0016] According to the generated label, the elements in the PQL statement are associated with the corresponding part of the interpretation text, thereby providing statement interpretation and partial highlight function, and the natural language interpretation of the PQL statement and the partial highlight relationship are displayed to the user.

[0017] S6, PQL statement fine-tuning:

[0018] According to the natural language interpretation of the PQL statement returned by the large language model, the user can send adjustment instructions to the large language model. Based on the PQL statement provided by the user and the adjustment instructions, the large language model can generate a fine-tuned PQL statement to meet the specific needs of the user.

[0019] Further, in step S1, the interpretation corresponding to the PQL statement is expressed in natural language.

[0020] Further, in step S1, the large language model is used to analyze the PQL statement structure relationship, and the specific meaning represented by the PQL statement is determined based on the analyzed PQL statement structure relationship. The specific meaning is expressed in natural language.

[0021] Further, in step S3, the PQL analysis prompt template is arranged based on the content of the natural language interpretation of the PQL statement according to the preset layout format, and the preset layout format is the PQL analysis prompt template.

[0022] Further, in step S5, based on the label name, the PQL statement elements associated with the label and their corresponding interpretation text content can be quickly determined.

[0023] Further, in step S5, the label name is the abbreviation of the PQL statement element associated with it by default.

[0024] Further, a system for applying a large language model analysis process to mine PQL, the system for applying a large language model analysis process to mine PQL internally carries a program and model for implementing steps S1-S5.

[0025] This invention provides a method and system for mining PQL based on a large language model parsing process, which has the following beneficial effects:

[0026] 1. This method and system for mining PQL based on a large language model parsing process uses a large language model to parse user-input PQL statements and convert them into easily understandable natural language text. This helps users quickly understand the specific meaning of PQL without human intervention. By providing highly readable natural language parsing results, it facilitates user reading and understanding while reducing the requirements for users to comprehend PQL statements. Attached Figure Description

[0027] Fig. 1 This is a schematic diagram illustrating the operation steps of a method for mining PQL based on a large language model parsing process according to the present invention.

[0028] Fig. 2 This is a schematic diagram of the PQL parsing process of the method for mining PQL based on a large language model parsing process according to the present invention. Detailed Implementation

[0029] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0030] like Figs. 1-2 As shown, the present invention provides a technical solution: a method for mining PQL based on a large language model parsing process. The method for mining PQL based on a large language model parsing process includes the following operations:

[0031] S1. Establish a large language model:

[0032] Prepare pre-defined PQL statements and their corresponding explanations, and use them as analysis examples to input into the large language model for sample learning and training, thereby confirming the structural relationships and output format of PQL statements.

[0033] The explanation of the PQL statement is expressed in natural language, while the large language model is used to parse the structural relationship of the PQL statement and determine the specific meaning of the PQL statement based on the parsed structural relationship. The specific meaning is expressed in natural language.

[0034] S2. Knowledge Base and Explanation Preparation:

[0035] Prepare the relevant structure and explanation of each element in PQL as a knowledge base to be used during parsing, so as to provide the specific meaning of the type parameters of each element in the PQL statement and the information of the corresponding fields;

[0036] S3 and PQL analysis tips:

[0037] In combination with the sample explanation and the PQL analysis prompt template, a PQL analysis prompt is generated, and the PQL statement provided by the user is transmitted to the large language model together with the PQL analysis prompt;

[0038] The PQL analysis prompt template is arranged according to a preset layout format based on the content of the natural language explanation of the PQL statement, and the preset layout format is the PQL analysis prompt template.

[0039] S4, PQL statement explanation:

[0040] The user selects a PQL statement and inputs it into the large language model. The large language model analyzes the PQL statement and determines the PQL statement structure relationship based on the analysis result of the PQL statement. The large language model searches the knowledge base based on the PQL statement structure relationship to return the natural language explanation of the PQL statement and the corresponding relationship label of the element and the explanation.

[0041] S5, partial highlight display:

[0042] According to the generated label, the elements in the PQL statement are associated with the corresponding part of the explanation text, thereby providing a statement explanation and partial highlight function. The natural language explanation of the PQL statement and the partial highlight relationship are displayed to the user.

[0043] The label name can quickly determine the PQL statement element associated with the label and its corresponding explanation text content. The label name is the abbreviation of the PQL statement element associated with it by default.

[0044] By using the large language model to analyze the PQL statement input by the user and converting it into natural language text that is easy to understand, the user can quickly understand the specific meaning of the PQL statement without human intervention. By feeding back the natural language analysis result with high readability, the user can easily read and understand the PQL statement while reducing the user's understanding requirements for the PQL statement.

[0045] S6, PQL statement fine-tuning:

[0046] According to the natural language explanation of the PQL statement returned by the large language model, the user can send adjustment instructions to the large language model. Based on the PQL statement provided by the user and the adjustment instructions, the large language model can generate a fine-tuned PQL statement to meet the specific needs of the user.

[0047] A system for applying a method of mining PQL in a large language model analysis process, which internally carries programs and models for implementing steps S1-S5.

[0048] In summary, as shown in the method and system for mining PQL based on a large language model analysis process, when used, first, a pre-set PQL statement and its corresponding explanation are prepared, which are used as analysis samples to input a large language model for sample learning and training, so as to confirm the PQL statement structure relationship and output format, and prepare the related structure and explanation of each element in PQL as a knowledge base used during analysis, to provide the specific meaning of each element type parameter of the PQL statement and the information of the corresponding field; Figs. 1-2

[0049] In combination with the sample explanation and PQL analysis prompt template, the PQL analysis prompt is generated, and the PQL statement provided by the user is transmitted to the large language model together with the PQL analysis prompt, and the PQL analysis prompt template is arranged according to the preset layout format based on the content of the natural language explanation of the PQL statement;

[0050] When the user end is operated, the user selects a PQL statement and inputs it to the large language model, the large language model analyzes the PQL statement, and determines the PQL statement structure relationship based on the analysis result of the PQL statement, the large language model searches the corresponding explanation in the knowledge base based on the PQL statement structure relationship, returns the natural language explanation of the PQL statement and the corresponding relationship label of the element and the explanation, according to the generated label, associates the element in the PQL statement and the corresponding part of the explanation text, so as to provide the statement explanation and local highlighting function, and then displays the natural language explanation and local highlighting relationship of the PQL statement to the user;

[0051] According to the natural language explanation of the PQL statement returned by the large language model, the user can send an adjustment instruction to the large language model, and based on the PQL statement provided by the user and the adjustment instruction, the large language model can generate a fine-tuned PQL statement to meet the specific needs of the user.

[0052] The embodiments of the present application are given for example and description, and are not exhaustive or limit the present application to the disclosed forms. Many modifications and changes are obvious to those skilled in the art. The embodiments are selected and described to better illustrate the principles and practical application of the present application, and to enable those skilled in the art to understand the present application so as to design various embodiments with various modifications suitable for specific purposes.​

Claims

1. A method for mining PQL based on large language model parsing process, characterized in that: The method for mining PQL based on a large language model analysis process comprises the following operations: S1, establishing a large language model: Prepare the preset PQL statement and its corresponding explanation as an analysis sample to input the large language model for sample learning and training, so as to confirm the PQL statement structure relationship and output format; S2, knowledge base and explanation preparation: Prepare the related structure and explanation of each element in PQL as the knowledge base used in analysis, so as to provide the specific meaning of each element type parameter in the PQL statement and the information of the corresponding field; S3, PQL analysis prompt: Generate a PQL analysis prompt by combining the sample explanation and the PQL analysis prompt template, and input the PQL statement provided by the user into the large language model together with the PQL analysis prompt; S4, PQL statement explanation: The user selects a PQL statement and inputs it into the large language model, the large language model analyzes the PQL statement, determines the PQL statement structure relationship based on the analysis result of the PQL statement, searches the corresponding explanation in the knowledge base based on the PQL statement structure relationship, and returns the natural language explanation of the PQL statement and the corresponding relationship label of the element and the explanation; S5, local highlight display: According to the generated label, the elements in the PQL statement are associated with the corresponding part of the explanation text, so as to provide the statement explanation and local highlight function, and the natural language explanation and local highlight relationship of the PQL statement are displayed to the user; S6, PQL statement fine-tuning: According to the natural language explanation of the PQL statement returned by the large language model, the user can send adjustment instructions to the large language model, and based on the PQL statement provided by the user and the adjustment instructions, the large language model can generate a fine-tuned PQL statement to meet the specific needs of the user.

2. The method for parsing a PQL based on a large language model according to claim 1, characterized in that: In the step S1, the explanation corresponding to the PQL statement is expressed in natural language.

3. The method of claim 1, wherein the method is based on a large language model parsing process to mine PQL. In the step S1, the large language model is used to analyze the PQL statement structure relationship, and the specific meaning represented by the PQL statement is determined based on the analyzed PQL statement structure relationship, which is expressed in natural language.

4. The method for mining PQL based on large language model parsing process according to claim 1, characterized in that: In the step S3, the PQL analysis prompt template is arranged based on the content of the natural language explanation of the PQL statement according to the preset layout format, that is, the PQL analysis prompt template.

5. The method of claim 1, wherein the method is based on a large language model parsing process to mine PQL. In the step S5, the label name can quickly determine the PQL statement element associated with the label and the corresponding explanation text content.

6. The method for parsing a PQL based on a large language model according to claim 1, characterized in that: In the step S5, the label name is the abbreviation of the PQL statement element associated with it by default.

7. A system for applying a method of mining PQL to a large language model parsing process, characterized in that: A method for mining PQL based on a large language model analysis process according to any one of claims 1-6, the system for mining PQL based on a large language model analysis process comprises programs and models for implementing steps S1-S5.

Images