A real-time text viewpoint extraction method based on syntax tree

Abstract

The invention discloses a real-time text viewpoint extraction method based on a syntax tree. The method comprises the following steps: a) inputting a text statement, wherein the text statement comprises a description word or an emotion word; B) splitting the text statement into sentences; (c) analyzing grammatical dependencies of each sentence to obtain a semantic relation graph, wherein the semantic relation graph is a directed acyclic graph, the vertex of the graph are words and parts of speech appearing in the sentence, and the edges of the graph are grammatical dependencies between words;D) searching for phrases of a particular pattern in the semantic relation graph; E) merging synonyms; f) counting phrase frequency; g) calculating phrase ranking; h) extracting the highest ranked viewpoint, and so on. The invention is different from the traditional algorithm, and can directly extract the important viewpoint in the comments, rather than a single word, and has important optimizationin computational efficiency and real-time performance.

Description

Technical field The invention relates to a real-time text opinion extraction method based on a syntax tree. Background technique With the development of the Internet, there are more and more online shopping platforms. Shopping platforms usually allow users to post their own comments on the products they have bought, and a large amount of comment data is accumulated over time. The mining of review text has important application value: for buyers, they can refer to other people's reviews to buy products that suit them; for sellers, they can refer to reviews to improve their product defects or improve service quality. However, it is time-consuming and laborious to read manually one by one in the face of massive review data. Therefore, it is necessary to extract the key points in the comments with the help of natural language processing algorithms. Traditional text key points extraction methods include: 1. Method based on word frequency and inverse text frequency. 2. The TextR...

AI Technical Summary

Problems solved by technology

These algorithms are more effective for extracting key points of traditional text content such as web pages and academic papers, but they are not ideal for online reviews, which are specifically reflected in the following aspects:

1. Emphasis on extracting single words, rather than a point of view with a complete meaning

2. Valuable negative views are easily buried

3. The calculation complexity is high, and it is not suitable for occasions with high real-time requirements

Images