Context-Based Carving of Fragmented Word Documents from Volatile Memory using Machine Learning Technique

Abstract

With the rise in digital crimes nowadays, digital investigators are required to recover and analyse data from various digital resources. Since the files are often stored in fragments owing to memory constraints, the information related to file system and metadata of the file is required to recover a file. However, in many cases when the file system is destroyed intentionally or unintentionally, and the metadata is deleted as well, the recovery of the digital evidence is done by a special method known as carving. In file carving, files are recovered solely based on the information about the structure and content of the individual file rather than matching the system’s information of the file.