An Enhanced Data Mining Classification based on Shuffled Frog Leaping Optimization

Abstract

Association Rule Mining (ARM) uncovers meaningful associations within discrete and categorical datasets. This paper introduces an innovative ARM method leveraging the Modified Shuffled Frog Leaping Optimization (MSFLO) technique to enhance performance analysis. By integrating the Apriori algorithm with MSFLO’s bio-inspired optimization, including frog encoding, our approach generates association rules efficiently. Unlike traditional methods requiring multiple database scans, this technique filters data in a single pass, significantly reducing CPU time and memory usage. Multiple optimization measures are applied to refine MSFLO, improving the accuracy and effectiveness of rule extraction. Implemented in MongoDB, the method is validated across six diverse datasets— Watermelon, Mangosteen, Breast, Dragon Fruit, Mango, and Orange—demonstrating superior performance compared to existing approaches. This advancement optimizes computational efficiency and rule quality, offering a robust solution for fruit shape database mining and precision agriculture applications.