Enhancing DDBMS Performance through RFO-SVM Optimized Data Fragmentation: A Strategic Approach to Machine Learning Enhanced Systems

Abstract

Effective data fragmentation is essential in enhancing the performance of distributed database management systems (DDBMS) by strategically dividing extensive databases into smaller fragments distributed across multiple nodes. This study emphasizes horizontal fragmentation and introduces an advanced machine learning algorithm, Red Fox Optimization-based Support Vector Machine (RFO-SVM), designed for optimizing the data fragmentation process. The input database undergoes meticulous pre-processing to address missing data concerns, followed by analysis through RFO-SVM. This algorithm efficiently classifies features and target labels based on class labels. The RFO algorithm optimizes critical SVM parameters, including the kernel, kernel parameter, and boundary parameter, leveraging the accuracy metric. The resulting classified data serves as fragments for the fragmentation process. To ensure precision in fragmentation, a Genetic Algorithm (GA) allocates these fragments to diverse nodes within the DDBMS, optimizing the total allocation cost as the fitness function. The proposed model, implemented in Python, significantly contributes to the efficient fragmentation and allocation of databases in distributed systems, thereby enhancing overall performance and scalability.