Extraction of Knowledge and Processing of the Patent Array

Abstract

In modern society, science and technology are developing with great speed. This leads to a massive accumulation of scientific and technical texts, their data structures are becoming more complex. There is a problem: how to deal with the growing volume of scientific and technical texts, as well as accurately determine their technical content, technical links, and trends. Traditional text analysis methods are based on the analysis of words and phrases to describe technological development (for example, the TFIDF method). One of the difficult problems that have arisen is how to integrate natural language processing, semantic analysis, and analysis of technology development trends, and how to perform a further in-depth study of the text to find hidden information about the development of technologies. Often, the development of technology is so fast that standardized technical terms are absent within even one technology, which makes it difficult to analyze textual information based on keywords. Therefore, in the framework of this study, it is proposed to use the machine learning method for SAO structures to solve the above problems. SAO (subject, action, object) structures are semantic structures that most accurately represent the semantic information of their textual data.