Enhanced Document Clustering for Forensic Analysis

Today, a crime can be in any form but the crime in digital form is increasing now a days. Therefore, the branch of computer science called digital forensic science has increasing importance. The forensic analysts analyze the computer data for particular proof against any crime. But, computers having huge amount of data files really creates havoc to analyze it. Therefore, it needs a good clustering techniques that reduces the efforts of forensic analysts. Already, there are many clustering techniques for the analysis purpose like (k-means, k-medoid, single link, complete link, etc. ) and found some good results. But, we used a new approach for the same purpose by applying k-representative as a key algorithm for clustering. We applied k-representative on the datasets collected from different sources and found really a good result as compared with the others. Also we focused on good preprocessing techniques like stemmer and porter algorithms. We experimented our techniques with different types of document formats and it works better with all formats. Finally, we shows results with comparative techniques with the help of comparative graphs.