Method for analyzing similarity of character string under Web environment

Abstract

The invention discloses a method for analyzing the similarity of character strings under a Web environment, comprising the following steps of: firstly, defining basic operation cost; pretreating the character strings to identify first characters thereof and remove the characters without real meaning; building distance matrix to realize the optimization of an edit distance by building a match index; then judging an abbreviated word, firstly judging whether the two character strings has an abbreviation relationship, if the abbreviation relationship is ensured, the distance is optimized; wherein, the abbreviation relationship is determined by two factors: (1) judging whether the two character strings has the similarity; and (2) judging whether the first characters of the two character strings have been matched; and after that optimizing the distance of the abbreviated word, wherein, the optimization is realized by reducing the cost for decreasing continuously inserted characters and continuously cancelled characters. The method for analyzing the similarity of the character strings can well treat the situations such as omission, abbreviation and character disorder which are usually in Web, has higher application, and under a Wed unknown environment has higher matching precision.

Description

technical field [0001] The invention belongs to the field of computer Web databases, and is particularly suitable for judging the similarity of two records in the repeated record recognition process of a Web database integration system. Background technique [0002] In the web environment, for strings that need to be matched by similarity, spelling mistakes, keyword order reversal, abbreviation or omitted word matching, etc. are often encountered, which leads to the string similarity analysis method applied in the web environment. Many difficulties. Because the typical commonly used string similarity analysis methods are usually only for a specific situation. For example: Levenshteindistance is more suitable for spelling mistakes, and Jaro distance metric is more suitable for abbreviation or omitted word recognition. In applications, it is often necessary to manually judge which algorithm to use in which environment. However, there are mostly semi-structured and unstructu...

AI Technical Summary

Problems solved by technology

However, this algorithm is only suitable for occasions where abbreviations are common. When comparing non-abbreviations, the similarity between the two words is often wrongly increased, resulting in distortion.

Images