A computer tool for data grouping based on behavior of bee

Abstract

In the field of data mining and unsupervised machine learning, data clustering is defined as the task of grouping objects according to a similarity or dissimilarity measure. That means, objects that are similar among them are grouped in the same cluster, and objects that are dissimilar are grouped into different clusters so a data descriptive structure can emerge. In social sciences, the classification and the grouping regarding to behavior patterns can take place to quantitative descriptions and predictions which let more specific study about how societies work under some parameters such as prediction of a crime emergent behavior in some social sectors. In general, the clustering problem can be formulated as a multi-objective optimization problem, which can be very complex in time and space computationally speaking. In this sense, the Artificial Bee Colony Algorithm which is a swarm intelligence algorithm based on numeric optimization, tries to get the best solution to the problem, exploiting and exploring the search space. In this work, we propose a computationally tool implemented in java for simulating the behavior of the honey bee swarms as a multi-agent system, where it is possible to observe the data clustering in training data that is used to tune the key parameters and compare them with similar papers. Through this experimentation, it is proposed to use the particle swarm optimization algorithm as a heuristic technique to get better initial solutions to the problem, so that the ABC algorithm can converge to a global optimum improving its convergence rate.