A method for creating structured documents based on deep learning models

Through the deep learning model-based method, the structure and logical relationship of text data are analyzed and the graph neural network model is constructed, which solves the problem of insufficient efficiency and accuracy of document structure construction in the existing technology, and realizes efficient and accurate structured document creation.

CN119558307BActive Publication Date: 2025-05-09FUJIAN WHALE AIR NETWORK TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202510115384.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-01-24
Publication Date
2025-05-09
Estimated Expiration
2045-01-24

AI Technical Summary

Technical Problem

When the existing technology processes massive and complex data, it is difficult to automatically build document structures, resulting in data omissions or inconsistencies, affecting the efficiency and accuracy of document creation.

Method used

Using a deep learning model-based method, we use a method to acquire and preprocess text data, analyze the structural and logical relationships of sentences and paragraphs, build a graph neural network model, and generate structured documents.

Benefits of technology

It realizes more accurate keyword extraction and paragraph core information recognition, improving the efficiency and accuracy of document creation, especially when processing complex documents, it can extract key semantic information.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119558307B_ABST
    Figure CN119558307B_ABST
Patent Text Reader

Abstract

The present invention provides a method for creating a structured document based on a deep learning model, and relates to the technical field of data processing. The method includes: obtaining basic text information collected by a user to obtain text data, slicing the text data to obtain text slicing data, analyzing the structure of sentences in a paragraph according to the text slicing data, determining keywords in the sentences, analyzing the logical relationship between sentences in the paragraph, associating the sentence keywords in the paragraph, analyzing the correlation between paragraphs in a chapter, identifying the core keywords of the chapter, obtaining a chapter keyword data set, analyzing the fields involved in the text, extracting the field information of the text, obtaining a text keyword data set, constructing a structured document model based on a graph neural network model, inputting the text keyword data set and the text data into the structured document model, and obtaining a structured document. The present invention can construct a structured document through a deep learning model.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Text semantic analysis method

    CN109271626A

  • Financial text summarization method based on keyword heterogeneous graph and semantic matching

    CN117725928A