An intelligent review method based on artificial intelligence and big data analysis

Abstract

The present invention discloses an intelligent review method based on artificial intelligence and big data analysis, which relates to a power system. Step one: data collection and preprocessing; Step two: based on the preprocessed data, model construction and training are performed to form a review model; Step three: based on the review model, intelligent review is performed; Step four: based on the review results, feedback and optimization are performed. In operation, the present invention uses information and digital means, combines the multi-dimensional and heterogeneous business data of the distribution network project, relies on an intelligent review model, and is performed by machine intelligent review, avoiding the disadvantages of relying on manual processing, which is cumbersome, has a large proportion of subjective factors, and has high business requirements for personnel.

Description

Technical Field

[0001] The present invention relates to a power system, and in particular to an intelligent review method based on artificial intelligence and big data analysis. Background Art

[0002] Traditional review processes in power systems rely heavily on manual labor, resulting in numerous issues such as low efficiency, high subjectivity, difficulty in standardizing standards, and chaotic data management during the review process. For example, in previous project reviews, manual record filling and report generation were not only time-consuming and labor-intensive, but also prone to errors. Different reviewers could have significantly different criteria for the same project, impacting the fairness and accuracy of the review results. With the rapid development of information technology, leveraging intelligent methods to optimize the review process and improve review quality has become a critical issue that needs to be addressed. Summary of the Invention

[0003] In response to the above problems, the present invention provides an intelligent review method based on artificial intelligence and big data analysis that saves manpower and improves efficiency.

[0004] The technical solution of the present invention is: an intelligent review method based on artificial intelligence and big data analysis, comprising the following steps:

[0005] Step 1: Data collection and preprocessing;

[0006] Step 2: Based on the pre-processed data, build and train the model to form an evaluation model;

[0007] Step 3: Intelligent review execution based on the review model;

[0008] Step 4: Provide feedback and optimization based on the review results.

[0009] Step one includes:

[0010] Obtain various data required for project review;

[0011] Clean the collected structured data to remove duplicate data, erroneous data, and incomplete data; identify unstructured data to form a structured data set;

[0012] The cleaned data is standardized and converted to meet the data format requirements within the system.

[0013] Step 2 includes:

[0014] Use natural language processing technology to conduct text mining on project application document data, extract key information and construct text feature vectors;

[0015] Combining historical review data and industry standard data, a review model is constructed using machine learning algorithms.

[0016] The review model develops corresponding review engines according to different project types. Each project type includes review type, review content category and review points.

[0017] Use cluster analysis algorithms to classify project data and identify characteristic patterns of different types of projects.

[0018] Step three includes:

[0019] When there is a new review project, the project data is input into the review model, and the review model automatically conducts a preliminary review and outputs preliminary review results and a list of issues;

[0020] For parts marked as requiring manual review in the preliminary review of the review model, the most suitable review expert is automatically matched based on the expert profile and project characteristics, and the task is assigned to the expert;

[0021] During the review process, experts use auxiliary tools to conduct review operations and directly fill in review opinions and problem details on the system interface. Among these tools, the auxiliary tools include the intelligent search function to quickly find relevant standards and historical cases.

[0022] The system records the experts’ operations and opinions in real time and integrates them with the preliminary review results of the model;

[0023] The final review report and statements are generated by integrating the review model and expert review opinions.

[0024] Step 4 includes:

[0025] Feedback the review results data, including project characteristics, review opinions, and final results, into model construction and training to update the review model and expert profiles;

[0026] Collect feedback from all parties during the review process, classify and analyze them, and optimize and improve the system's review standards based on the analysis results.

[0027] In operation, the present invention uses information and digital means, combines the multi-dimensional and heterogeneous business data of the distribution network project, relies on an intelligent review model, and is executed by machine intelligent review, avoiding the disadvantages of relying on manual processing, which has a cumbersome process, a large number of subjective factors, and high business requirements for personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the specific embodiments or the description of the prior art. In the drawings, the various parts are not necessarily drawn according to the actual scale.

[0029] Figure 1 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0030] The present invention Figure 1 As shown, an intelligent review method based on artificial intelligence and big data analysis is provided, which includes the following steps:

[0031] Step 1: Data collection and preprocessing;

[0032] Step 2: Based on the pre-processed data, build and train the model to form an evaluation model;

[0033] Step 3: Intelligent review execution based on the review model;

[0034] Step 4: Provide feedback and optimization based on the review results.

[0035] In operation, the present invention uses information and digital means, combines the multi-dimensional and heterogeneous business data of the distribution network project, relies on an intelligent review model, and is executed by machine intelligent review, avoiding the disadvantages of relying on manual processing, which has a cumbersome process, a large number of subjective factors, and high business requirements for personnel.

[0036] Specifically,

[0037] Step 1: Data collection and preprocessing

[0038] ① Connect multiple data sources and, according to pre-defined data collection rules, obtain the various data required for project review. For example, basic project information and related attachments, including 11 types of documents: project feasibility study report, project budget, project design plan, design drawings, grid power supply reliability impact assessment report, supply list (Part A), supply list (Part B), etc. The required formats include World Wide Web, Excel, PDF, CAD, and images.

[0039] ② Clean the collected structured data to remove duplicate data, erroneous data, and incomplete data; perform intelligent identification on unstructured data (various attachments) to form a structured data set.

[0040] ③ Standardize and convert the cleaned data to meet the data format requirements within the system.

[0041] Step 2: Model building and training

[0042] ① Use natural language processing technology to conduct text mining on project application document data, extract key information and construct text feature vectors.

[0043] ② Combining historical review data with industry standard data, a machine learning algorithm was used to construct a review model. This model developed a specific review engine based on different project types: grid projects, non-grid (capital) projects, non-grid (cost) projects, and rural power grid projects. Each project type involves three major review types, 14 review content categories, and 76 review points, as shown in Table 1.

[0044]

[0045]

[0046] Table 1

[0047] ③ Use cluster analysis algorithms to classify project data, identify characteristic patterns of different types of projects, and provide a basis for subsequent personalized reviews.

[0048] Step 3: Intelligent review execution

[0049] ① When a new project is reviewed, the project data is fed into the trained review model, which automatically conducts a preliminary review and outputs preliminary review results and a list of issues. For example, based on the project type (gridframe or non-gridframe), the model determines whether the project's supporting attachments comply with regulations, whether the project's basic information is consistent with the supporting materials, and whether the project's plan is reasonable.

[0050] ② For the parts marked as requiring manual review in the initial model review, the most suitable review expert is automatically matched based on the expert profile and project characteristics, and the task is assigned to the expert.

[0051] Expert Allocation Principles: An automatic expert allocation algorithm is developed based on multiple factors, including the expert's professional expertise (first-level and second-level categories), review direction, professional capabilities, work performance, unit, and project review points. The algorithm prioritizes the mutual exclusion principle between the expert's unit and the project unit.

[0052] ③ During the review process, experts use the system's auxiliary tools to conduct review operations, such as using the intelligent search function to quickly find relevant standards and historical cases, and directly fill in review comments and detailed questions on the system interface. The system records the experts' operations and opinions in real time and integrates them with the model's initial review results.

[0053] ④ The final review report and report are generated based on the integrated model and expert review opinions. The review report includes the overall review conclusions of the project, detailed problem analysis, and corrective suggestions. The review report displays key review information of the project in a pre-set format, such as the review score and the classification and statistics of major problems.

[0054] Step 4: Feedback and Optimization

[0055] ① Feedback the review results, including project characteristics, review opinions, and final results, into model construction and training to update the review model and expert profiles. For example, if a new project feature is found to have a significant impact on the review results but was not fully considered in the previous model, this feature will be added to the model's training data and the model will be retrained to improve its accuracy.

[0056] ② Gather feedback from all parties involved in the review process, such as questions from project applicants regarding the review criteria and suggestions from review experts regarding system functionality. This feedback is categorized, organized, and analyzed, and based on the analysis, improvements are made to the system's review criteria and functional modules. For example, if multiple experts report that the system's intelligent notification feature is inaccurate, the notification algorithm will be optimized to improve the relevance and accuracy of the notification information.

[0057] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. An intelligent review method based on artificial intelligence and big data analysis, characterized in that: The following steps are involved: Step 1: Data collection and preprocessing, including multi-dimensional and heterogeneous business data of distribution network projects; Step 2: Based on the pre-processed data, build and train the model to form an evaluation model; Step 2 includes: Use natural language processing technology to conduct text mining on project application document data, extract key information and construct text feature vectors; Combining historical review data and industry standard data, we use machine learning algorithms to build review models; Step 3: Intelligent review execution based on the review model; Step three includes: When there is a new review project, the project data is input into the review model, and the review model automatically conducts a preliminary review and outputs preliminary review results and a list of issues; For parts marked as requiring manual review in the preliminary review of the review model, the most suitable review expert is automatically matched based on the expert profile and project characteristics, and the task is assigned to the expert; During the review process, experts use auxiliary tools to conduct review operations and directly fill in review opinions and problem details on the system interface. Among these tools, the auxiliary tools include the intelligent search function to quickly find relevant standards and historical cases. The system records the experts’ operations and opinions in real time and integrates them with the preliminary review results of the model; Comprehensively combine the review model and expert review opinions to generate the final review report and statements; Step 4: Provide feedback and optimization based on the review results; Step 4 includes: Feedback the review results data, including project characteristics, review opinions, and final results, into model construction and training to update the review model and expert profiles; Collect feedback from all parties during the review process, classify and analyze it, and optimize and improve the system's review standards based on the analysis results; The review model develops corresponding review engines according to different project types. Each project type includes review type, review content category and review points; Use cluster analysis algorithms to classify project data and identify characteristic patterns of different types of projects.

2. The intelligent review method based on artificial intelligence and big data analysis according to claim 1 is characterized in that: Step one includes: Obtain various data required for project review; Clean the collected structured data to remove duplicate data, erroneous data, and incomplete data; identify unstructured data to form a structured data set; The cleaned data is standardized and converted to meet the data format requirements within the system.

Images