Investigation on Digital Forensic Using Graph Based Neural Network With Blockchain Technology

Digital forensics has evolved to focus on leveraging social media evidence, offering substantial support for investigating various crimes. However, scrutinizing social media data for evidence presents significant challenges for authorities. This article proposes employing blockchain infrastructure and natural language processing (NLP) techniques as foundational elements in digital forensic investigations. Specifically, it advocates for a Graph Convolutional Neural Network (GCNN) based approach to multi-class classification. NLP is integral to various stages of this process, including data collection, feature selection, classifier assessment, and vectorization. Moreover, to enhance data security against hackers and network attacks, a blockchain-based system is utilized. Real-world dataset experimentation demonstrates the efficacy of the proposed system. The findings reveal significant improvements in GCNN classification performance when incorporating topology knowledge. Furthermore, optimizing the model necessitates considering various input factors, such as document length, training data representation, and network architecture. This structured approach to digital forensic investigation holds promise for bolstering law enforcement efforts.