New Descriptors of Textual Records: Getting Help from Frequent Itemsets

Abstract

The analysis of numerical data, whether structured, semi-structured, or raw, is of paramount importance in many sectors of economic, scientific, or simply social activity. The process of extraction of association rules is based on the lexical quality of the text and on the minimum support set by the user. In this paper, we implemented a platform named “IDETEX” capable of extracting itemsets from textual data and using it for the experimentation in different types of clustering methods, such as [Formula: see text]-Medoids and Hierarchical clustering. The experiments conducted demonstrate the potential of the proposed approach for defining similarity between segments.