Text correlation detection method, device, equipment and storage medium

Abstract

This application relates to the technical field of speech semantics, and discloses a text correlation detection method, device, device, and storage medium. The method includes: forming a first keyword vector and a second keyword vector, and analyzing the first keyword vector based on a fuzzy mean algorithm. The keyword vector and the second keyword vector carry out cluster analysis to obtain a theme vector; determine the first theme of the first text and the second theme of the second text based on the theme vector, and construct a theme map according to the first theme and the second theme; Based on the preset keyword matching rules, the first matching feature and the second matching feature are constructed for the sentence pairs under each topic node, and the first matching feature and the second matching feature of any topic node are spliced ​​to obtain the topic node. Matching feature: input the matching feature and the connection matrix into the graph convolution model for analysis, and obtain the correlation between the first text and the second text. The global matching between texts is realized, and the flexibility and accuracy of correlation matching are improved.

Description

technical field [0001] The present application relates to the technical field of speech semantics, in particular to a text correlation detection method, device, equipment and storage medium. Background technique [0002] Text matching is to calculate the correlation between two texts. At present, text matching is often used to check the correlation between the business text provided by the staff and the industry standard text, so as to determine the business ability of the corresponding staff. For example, the common testing process of customer service speech quality is the process of correlating the call text between customer service and users with industry standard speech. Traditional detection methods are mainly rule-based face-to-face notification matching or tf-idf-based matching, but the rule-based face-to-face notification matching method lacks flexibility and incompleteness, and the matching algorithm based on tf-idf can only achieve accurate matching. It is diffic...

AI Technical Summary

Problems solved by technology

Traditional detection methods are mainly rule-based face-to-face notification matching or tf-idf-based matching, but the rule-based face-to-face notification matching method lacks flexibility and incompleteness, and the matching algorithm based on tf-idf can only achieve accurate matching. It is difficult to accurately assess the degree of similarity for texts with similar speech and large differences in glyphs

[0003] To sum up, the existing text correlation matching methods lack flexibility, and there is a problem that global matching cannot be performed

Images