Scientific and technological paper data text semantic feature extraction method and system and storage medium

Abstract

The invention provides a science and technology paper data text semantic feature extraction method and system and a storage medium, and the method comprises the steps: obtaining the text information of a science and technology paper, and constructing an entity relation graph based on the obtained text information of the science and technology paper, the text information comprises a paper title and a keyword, and the entity relation graph comprises the keyword; nodes in the entity relationship graph are paper titles or keywords, and edges in the entity relationship graph are association relationships among the nodes; semantic features are extracted based on the obtained text information of the science and technology papers, and a semantic feature matrix is obtained; determining an original adjacency matrix based on the entity relation graph, and inputting the semantic feature matrix and the original adjacency matrix into a graph network model to obtain a spatial feature matrix; and carrying out feature fusion on the semantic feature matrix and the spatial feature matrix to obtain final semantic features of the science and technology paper. According to the feature extraction method, on the basis of extracting the semantic features of the scientific and technological paper corpus, the semantic features of the scientific and technological paper can be well extracted by utilizing the spatial association of the knowledge graph.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a method, system and storage medium for extracting semantic features of scientific and technological paper data text. Background technique [0002] As an important source of research results display and information acquisition, a large number of scientific papers are published almost every day. These academic results contain a variety of the latest professional field information. These scientific papers can be obtained effectively and quickly and the semantic features can be represented. and learning is particularly important. However, the data of scientific papers often contain a large number of complex attributes, such as abstracts, keywords, citations, etc. of the papers, and the correlation between papers is closer. In addition, a large amount of professional knowledge in papers covers a wide range of disciplines, which makes the feature extraction of scientific papers It ...

AI Technical Summary

Problems solved by technology

[0005] In the existing feature extraction methods based on the pre-trained corpus model, the semantic feature extraction is carried out for the text context. However, for scientific and technological papers, various attributes of the paper, especially keywords and titles, have a large number of correlations. These attributes can also be associated through the co-occurrence relationship of keywords. These attributes cover the main semantic information of the paper, but the existing feature extraction methods ignore the association between these paper attributes, and cannot be analyzed from the context and the paper association relationship at the same time. Feature extraction, which leads to the inability of existing feature extraction methods to extract the semantic features of scientific papers

Images