Search

CN-122021575-A - Method and device for filling project item information and electronic equipment

CN122021575ACN 122021575 ACN122021575 ACN 122021575ACN-122021575-A

Abstract

The application relates to the technical field of data processing and discloses a method for filling project stand information, which comprises the steps of analyzing a project stand document by utilizing a semantic understanding model to identify key fields; and filling the field contents in the unified data format into a form of the target system according to the system form field. According to the method, the file analysis and the field mapping are realized through the AI model, so that the time of manual input is obviously reduced, and the project standing efficiency is improved. And analyzing the project stand documents through the semantic understanding model, and processing the documents with different structures, so that the accuracy of filling the project stand information is improved. The application also discloses a device for filling project item information and electronic equipment.

Inventors

  • SHI LU
  • CUI XIUYUAN
  • ZHAO HUAWEI

Assignees

  • 青岛巨商汇网络科技有限公司

Dates

Publication Date
20260512
Application Date
20251230

Claims (10)

  1. 1. A method for populating project legislation information, comprising: analyzing the project stand document by utilizing the semantic understanding model to identify key fields; mapping the key fields to system form fields of the project management system, and generating field contents in a unified data format; And filling field contents in a unified data format into a form of the target system according to the system form fields.
  2. 2. The method of claim 1, wherein parsing the project legislation document using the semantic understanding model to identify key fields comprises: Carrying out structural analysis on the project standing document to obtain an analysis result; and carrying out semantic understanding on the analysis result by utilizing a semantic understanding model and identifying key fields.
  3. 3. The method of claim 2, wherein structurally parsing the project legislated document to obtain parsing results comprises: extracting information data in the project setting documents, wherein the information data comprises graphic data and chart data; Mapping information data in the project standing document to generate a node tree; and recording node information of each node in the node tree, and taking the node information as an analysis result of the project standing document.
  4. 4. A method according to claim 3, wherein semantically understanding the parsing result and identifying key fields using a semantic understanding model comprises: Inputting the image-text data into a transducer model to obtain the key field of the image-text data output by the transducer model and the confidence corresponding to the key field; And extracting node sequence, hierarchy, time axis scale and numerical mapping relation of SmartArt nodes in the chart data so as to carry out structural processing on the chart data.
  5. 5. The method of claim 4, wherein inputting the teletext data into the transducer model to obtain the key fields and the confidence levels corresponding to the key fields of the teletext data output by the transducer model comprises: image preprocessing is carried out on the image-text data, and image-text characteristics in the image-text data are extracted, wherein the image-text characteristics comprise visual characteristics and text characteristics; The visual characteristics and the text characteristics are spliced and then input into a Transformer encoder, and the image-text characteristics are subjected to iterative optimization through the Transformer encoder to obtain fusion characteristics; The fusion features are converted into structured fields containing confidence and serve as key fields for the project legislation document.
  6. 6. The method of any of claims 1 to 5, wherein mapping key fields to system form fields of the project management system and generating field contents in a unified data format comprises: Performing synonymous clustering operation through a graph neural network to obtain a key field relation graph, wherein the key field relation graph comprises the business attributes of fields; And carrying out synonym mapping on the key fields according to the key field relation graph, and generating field contents in a unified data format.
  7. 7. The method according to any one of claims 1 to 5, further comprising: Filling field contents in a unified data format into a form of a target system, and feeding back filling results to a user; obtaining a verification result of a filling result by a user; and optimizing the semantic understanding model according to the verification result.
  8. 8. An apparatus for populating project legislation information, comprising: The document analysis module is configured to analyze the project stand document to identify key fields; the field mapping module is configured to map the key fields to system form fields of the project management system and generate field contents in a unified data format; And the form filling module is configured to fill field contents of the unified data format into a form of the target system according to the system form fields.
  9. 9. An apparatus for populating project subject information comprising a processor and a memory storing program instructions, wherein the processor is configured, when executing the program instructions, to perform the method for populating project subject information of any one of claims 1 to 7.
  10. 10. An electronic device, comprising: An electronic device body; the apparatus for filling item legislation information of claim 8 or 9, mounted to the electronic device body.

Description

Method and device for filling project item information and electronic equipment Technical Field The present application relates to the field of data processing technologies, and for example, to a method and an apparatus for filling project item information, and an electronic device. Background At present, in the IT project setting process, a large number of repeated form fields need to be filled manually, data formats among different systems are different, project declaration personnel need to switch among multiple systems repeatedly, and therefore the efficiency of filling project setting information is low. In order to improve the efficiency of filling project item information, the information processing and form filling method disclosed by the related technology comprises the steps of analyzing a document by adopting an OCR (optical character recognition) character recognition technology and a keyword matching extraction technology, processing natural language based on a document abstract generation technology and an information extraction system, and combining a manual filling system based on a template with a browser automatic filling plug-in. In the implementation of the embodiment of the disclosure, the related technology is found to have at least the following problems that although the related technology can reduce local manual input operation and extract basic text information of a document, the related technology still cannot process unstructured documents (such as PPT stand materials containing graphic-text mixed arrangement and SmartArt charts), the fact that semantic association among information leads to low extraction accuracy of key information, poor semantic recognition effect and incapability of realizing automatic filling of information of IT projects. It should be noted that the information disclosed in the above background section is only for enhancing understanding of the background of the application and thus may include information that does not form the prior art that is already known to those of ordinary skill in the art. Disclosure of Invention The following presents a simplified summary in order to provide a basic understanding of some aspects of the disclosed embodiments. This summary is not an extensive overview, and is intended to neither identify key/critical elements nor delineate the scope of such embodiments, but is intended as a prelude to the more detailed description that follows. The embodiment of the disclosure provides a method and a device for filling project stand information, and electronic equipment, and can improve the accuracy of filling the project stand information. In some embodiments, the method for populating project legislation information comprises parsing the project legislation document with a semantic understanding model to identify key fields, mapping the key fields to system form fields of a project management system and generating field contents in a unified data format, and populating the field contents in the unified data format into a form of a target system according to the system form fields. Optionally, the project stand document is analyzed by using a semantic understanding model to identify key fields, including structural analysis of the project stand document to obtain an analysis result, and semantic understanding of the analysis result and identification of the key fields are performed by using the semantic understanding model. Optionally, carrying out structural analysis on the project stand document to obtain an analysis result, wherein the analysis result comprises the steps of extracting information data in the project stand document, wherein the information data comprises graphic data and chart data, carrying out mapping processing on the information data in the project stand document to generate a node tree, recording node information of each node in the node tree, and taking the node information as the analysis result of the project stand document. Optionally, semantic understanding is carried out on the analysis result by utilizing a semantic understanding model, and key fields are identified, wherein the semantic understanding comprises the steps of inputting image-text data into a transducer model, obtaining the key fields of the image-text data output by the transducer model and the confidence corresponding to the key fields, and extracting node sequence, hierarchy, time axis scale and numerical mapping relation of SmartArt nodes in chart data so as to carry out structural processing on the chart data. Optionally, inputting the image-text data into a transducer model to obtain the confidence coefficient corresponding to the key field and the key field of the image-text data output by the transducer model, wherein the image-text data is subjected to image preprocessing to extract image-text features in the image-text data, the image-text features comprise visual features and text features, the visual features and the text fea