A Neural Network based Digital Forensics Classification

Digital forensics (DF) turn out to be an urgent and a timely subject to investigate with the significant rise in computer crimes these days. Typically, DF analysis aims to maintain all collected evidence in its most original status by determining, collecting, and evaluating the digital data for re-constructing past incidents. Most evidence related to digital crime are kept within the computer system files. This article explores and evaluates the applicability of Neural Network techniques in DF analysis by analysing information related to computer's file system to determine whether they have been manipulated by a specific application program. A data set described as a vector of attributes related to file system activities thru a specific time is collected and utilized for creating a neural network classification model. The experimental results show good results with respect to different performance estimation measures.