Short Text Feature Extension Method Based on Semantic Graph

Abstract

The invention discloses a short text feature extension method based on a semantic map, which includes the following steps: using a short text training data set to perform topic modeling, extracting the distribution of topic words; reordering the distribution of topic words; constructing a dictionary of candidate keywords and Topic-keyword semantic map; based on the method of link analysis, the comprehensive similarity evaluation of candidate keywords and seed keywords is calculated, and the most similar candidate keywords are selected to complete the expansion of the short text. Compared with the short text feature representation method based on the language model, the method of the present invention is simple to operate, has high execution efficiency, and makes full use of the semantic correlation information between keywords, compared with the traditional short text feature representation method based on the bag of words model, effectively alleviating Sparsity issues and semantic sensitivity issues do not depend on external large-scale auxiliary training corpus or search engines.

Description

technical field [0001] The present invention relates to the technical field of text mining, and is a short text feature extension method based on topic-keyword semantic map and link analysis, which can be applied to feature representation in short text classification and clustering tasks, and finally applied to knowledge question answering, Subfields such as user intent understanding and intelligent retrieval. Background technique [0002] With the advent of the era of big data, the Internet and various mobile terminals have generated a large amount of short text information, such as web page retrieval fragments, Weibo, product reviews, news headlines, and various micro-information, etc. Information is also being overwhelmed by the vast amount of resources. How to make the system intelligently manage and better use these massive data resources is facing a huge challenge. Therefore, a high-precision short text classification method can help the system to deepen the understa...

AI Technical Summary

Problems solved by technology

[0006] Aiming at the above two main problems, the present invention proposes a short text feature extension method based on semantic maps, which solves the problem of data sparsity and semantic sensitivity in the representation of short text features by the traditional bag-of-words model, and finally improves short text features. Ben's classification performance

Images