Deceptive Operations in Document Repositories: Manipulating Clustering Outcomes Against Adversaries

Abstract

This research investigates the efficacy of replacement and shuffling techniques to enhance the confidentiality and integrity of sensitive information within diverse document types. The study introduces the Deceptive Approaches for Robust Defense (DARD) technique, which aims to anonymize and protect numerical data and confidential text. The effectiveness of this technique is evaluated using three distinct datasets. The first dataset consists of 300 research papers on Artificial Intelligence, Cryptography, and Databases. The second dataset includes summaries of 3000 research papers spanning Artificial Intelligence, Cryptography, Databases, and Networking. The third dataset encompasses company documents classified into Inventory Reports, Invoices, Purchase Orders, and Shipping Orders. The comparative analysis between the first and second datasets, and the first and third datasets, demonstrates the DARD technique’s proficiency in anonymizing and securing sensitive data across various document types. The findings reveal that the DARD technique effectively safeguards confidential information in both academic research papers and business documents, with a particular strength in handling documents containing numerical data and sensitive content. This research contributes to the field of data security by providing a robust method for protecting sensitive documents, thereby addressing critical issues in cybersecurity practices. The study underscores the potential of the DARD technique to serve as a reliable tool for ensuring data confidentiality and integrity, offering significant implications for both academic and commercial applications. The results validate the technique’s applicability in real-world scenarios, highlighting its importance in the ongoing efforts to enhance data privacy and security