A Technique for Discovering Similarities between Texts Based on Extracting Features from the Text

Abstract

The discovery of the similarity between two texts is very important and useful in many applications. The similarity between texts is the core research area of dataset, data warehouse, and data mining. This paper provides a framework that gives a similarity between two input texts based on pattern recognition and the use of approximate string matching; there is a weight that affects the proportion of similarity. The search compares the similarity of two texts without adherence to the grammar or the use of synonyms or meanings of words. Preliminary results showed the benefit of extracting some of the features in the discovery of the similarity between the texts.