Text Document Clustering Using Dimension Reduction Technique

Text document clustering is used to group a set of documents based on the information it contains and to provide retrieval results when a user browses the internet. Experimental evidences have shown that Information Retrieval applications can benefit from document clustering and it has been used as a tool to improve the performance of retrieval of information. Information retrieval is an interdisciplinary field of knowledge management and text mining. Dimensionality Reduction (DR) is a typical step in many text mining problems which involves transforming sparse data into a shorter and more compact one. DR can be done in 2 ways: feature reduction and feature selection. This study implements dimensionality reduction through feature selection with k-means algorithm. Feature Selection is implemented through the InfoGain DR technique. This paper presents an experimental analysis of the performance of the document clustering with the InfoGain technique and proves that this method significantly improves the performance in terms of Accuracy, Precision and Recall for the BBC Sports Dataset.