Artificial Bee Colony Algorithm and Its Application to Content Filtering in Digital Communication

Abstract

In nature, some species of creatures form swarms to perform their daily tasks such as guarding themselves, foraging and group decision making. An intelligent swarm performs its tasks by optimizing the sources and responding to the environmental changes adaptively. One of the intelligent swarms is a honeybee colony in which the tasks are assigned to the bees in the swarm according to the hive conditions. The task division and self-organization skills of honeybees lead to swarm intelligence. Artificial bee colony (ABC) models the intelligent foraging behavior of a honeybee colony in which the honey unloaded to the hive is maximized by self-organization without a supervision. In this chapter, first, the collective intelligence arising in the foraging of a real honeybee is described, and then the basic components of ABC are explained. An application is also provided to show how ABC can be used for automated filtering unsolicited digital content in which robust classifiers trained by efficient algorithms are needed. In the application, each content is represented by features based on mutual information and $tf-idf$ metrics, and a logistic regression classifier trained by ABC algorithm classifies the content efficiently in terms of classification accuracy.