Shakir Syed

The advent of big data has had a significant impact on enterprise resource planning (ERP) ecosystems, particularly when it comes to supporting scalability and addressing the limitations of existing cybersecurity frameworks in ERP ecosystems. Big data technologies enhance cybersecurity in ERP ecosystems by improving cyber forensics readiness. However, the use of big data-enhanced cybersecurity solutions in ERP ecosystems can result in several cybersecurity concerns regarding the privacy, protection, and preservation of ERP cybersecurity data. This creates a need for machine learning-based algorithms to optimize data analytics-based cybersecurity initiatives in ERP ecosystems. Five machine learning algorithms are developed to optimize ERP big data-enhanced cybersecurity. In developing these machine learning models, a framework has been created for securing ERP big data by identifying and selecting the most appropriate machine learning algorithm that can be utilized to develop an effective ERP cybersecurity solution. Each of the algorithms has been analyzed in terms of its evaluation metrics and other performance and learning attributes. While all the algorithms can be effectively used for ERP big data-enhanced cybersecurity, the following are the outstanding strengths of each algorithm: logistic regression for ensuring scalability and making classification predictions based on real-time assessments; decision tree for easily integrating with existing ERP systems; random forest for its ensemble learning-based power to enhance overall ERP ecosystem security; k-nearest neighbors for its simple and easy-to-understand methodology; and support vector machine for its potential in ERP system security data clustering and addressing the system security multi-dimensionality challenge. This paper also presents the limitations of each evaluated algorithm, as well as areas for additional research to maximize the overall effectiveness of the machine learning techniques.