Keyword weight calculation method and device, computer equipment and storage medium

Abstract

The embodiment of the invention discloses a keyword weight calculation method and device, computer equipment and a storage medium. The method comprises the steps of obtaining query content input by a user; performing word segmentation processing on the query content to obtain keywords corresponding to the query content; inputting the query content into the trained BERT network model for bidirectional language representation processing to obtain a word vector of each word in the query content; determining a keyword vector corresponding to the keyword according to the word vector corresponding to each word in the keyword; and inputting the keyword vector into the trained feedforward neural network model to obtain a weight corresponding to the keyword. In the embodiment of the invention, the obtained keyword vector is related to the context in the query content and is the dynamic keyword vector, then the weight of the dynamic keyword vector is predicted according to the feed-forward neural network to obtain the dynamic weight of the keyword, and when the dynamic weight is applied to a search engine, the recall rate of the search engine can be improved.

Description

technical field [0001] The present invention relates to the technical field of the Internet, in particular to a keyword weight calculation method, device, computer equipment and storage medium. Background technique [0002] With the development of the Internet, the amount of stored data in the Internet is very large. Therefore, in order to enable users to quickly and accurately find the data content they need, manufacturers that provide Internet search services need to optimize the search quality of search engines. Among them, the keyword weight is an evaluation value given by the search engine to the search content. This weight can reflect the importance of the keyword. The higher the weight, the more important the keyword is. [0003] In the process of using the search engine, the user will submit the query content in the search box. These query content are usually called query. The search engine needs to obtain useful information from the massive data according to the que...

AI Technical Summary

Problems solved by technology

[0004] In the prior art, static weight algorithms such as Term Frequency–Inverse Document Frequency (TF-IDF) or BM25 based on statistics are generally used to calculate the weight of keywords. Among these static weight algorithms , each keyword has a preset corresponding weight, because the same keyword in different contexts, its weight may be different, if you use a fixed weight, it will affect the search engine's recall effect

Images