Automatic error correction method for Chinese search term of search engine

Abstract

The invention relates to an automatic error correction method for a Chinese search term of a search engine and belongs to the technical field of computer applications. According to the automatic error correction method, a data module, an offline database construction end and an online retrieval end are included, wherein the data module is mainly used for providing data for the offline database construction end and the online retrieval end. Abundant offline data are mined by using modules, such as a search log, a crawler system and the like and are used for various error correction strategies. For different fields, a dictionary in an exclusive field is used for system error correction. According to the automatic error correction method for the Chinese search term of the search engine, a manner of combining various independent error correction strategies is adopted, various strategies supplement and compare with each other for a complex query error, and a good result is finally achieved. As the error correction for a second time is employed, an error correction range is wide and the success rate is high. The error correction strategies can be flexibly and independently configured, an error correction occasion is wide, and the automatic error correction method can adapt to various different vertical search fields.

Description

technical field [0001] The invention relates to an automatic error correction method for Chinese search words of a search engine, and belongs to the technical field of computer applications. Background technique [0002] Today, search engines are one of the most important ways for people to obtain information. The most basic and core function of a search engine system is information retrieval, finding webpages or documents containing keywords, and then giving the results according to a certain order. In the search engine, we call the keyword information entered by the user query, and the user hopes to get a web page or document with better quality related to the input query. There are many ways to measure the word "good", and the simplest standard is those The most helpful and attractive results to users can be ranked first. However, due to various reasons, the quality of the query itself entered by the user is not high or wrong. If the search engine does not correct and m...

AI Technical Summary

Problems solved by technology

Chinese characters also have problems similar to Real-word Error, such as salary increase imperial decree, salary increase and imperial decree are both correct words, but there is indeed a problem with the two together, so in many cases Chinese query error correction is actually a phrase error correction problem

[0007] 2. Query error correction technology is complicated:

In addition, since Chinese words are often short, the difference of one word may completely change the meaning of the word, so the single use of error correction methods such as edit distance often leads to a greater risk of escaping

Images