Global Artificial Bee Colony Search Algorithm for Data Clustering

Abstract

Data clustering is a widespread data compression, vector quantization, data analysis, and data mining technique. In this work, a modified form of ABC, i.e. global artificial bee colony search algorithm (GABCS) is applied to data clustering. In GABCS the modification is due to the fact that experienced bees can use past information of quantity of food and position to adjust their movements in a search space. Due to this fact, solution search equations of the canonical ABC are modified in GABCS and applied to three famous real datasets in this work i.e. iris, thyroid, wine, accessed from the UCI database for the purpose of data clustering and results were compared with few other stated algorithms such as K-NM-PSO, TS, ACO, GA, SA and ABC. The results show that while calculating intra-clustering distances and computation time on all three real datasets, the proposed GABCS algorithm gives far better performance than other algorithms whereas calculating computation numbers it performs adequately as compared to typical ABC.

Association Rule Mining Based on Hybrid Whale Optimization Algorithm

“…Artificial Bee Colony Algorithm (ABC) had a good balance between intensification and diversity (Akay et al, 2021). Danish et al (2019) proposed a global ABC (GABCS) for data clustering. Sharma et al (2015) used ABC to generate high-quality association rules for searching frequent itemsets from large data sets.…”