This study addresses the escalating volume of research by proposing an efficient research storage system through data mining-based categorization. Employing the Support Vector Machine (SVM) method on a dataset comprising 541,910 retail product purchases, the research achieves a significant 96.2% accuracy in categorization using the cross-entropy loss function. The SVM method proves instrumental in systematically organizing research based on fields, methods, and outcomes, showcasing its efficacy in large-scale research storage and organization. This study highlights the SVM's potential as a vital tool for governments and private organizations to enhance access and utilization of research information. The results underscore the positive impact of SVM in overcoming the complexity of research storage on a broader scale, contributing to the advancement of efficient research management systems.
Highlights:
Efficient SVM Data Management: Proposes SVM-based data mining for effective research information storage.
96.2% Accuracy in Categorization: SVM with cross entropy achieves high accuracy in classifying research data.
Organized Access for Better Utilization: SVM organizes research systematically, enhancing accessibility and utilization for government and private sectors.
Keywords: Support Vector Machine, Data Mining, Dataset, Retail.