Ant Colony Optimization and Data Mining

The Ant Colony Optimization (ACO) technique was inspired by the ants' behavior throughout their exploration for food. In nature, ants wander randomly, seeking for food. After succeeding, they return to their nest. During their move, they lay down pheromone that forms an evaporating chemical path. Other ants that locate this trail, follow it and reinforce it, since they also lay down pheromone. As a result, shorter paths to food have more pheromone and are more likely to be followed. ACO algorithms are probabilistic techniques for solving computational problems that are based in finding as good as possible paths through graphs by imitating the ants’ search for food. The use of such techniques has been very successful for several problems. Besides, Data Mining (DM), a discipline that consists of techniques for discovering previously unknown, valid patterns and relationships in large data sets, has emerged as an important technology with numerous practical applications, due to wide availability of a vast amount of data. The collaborative use of ACO and DM (the use of ACO algorithms for DM tasks) is a very promising direction. In this chapter, we review ACO, DM, Classification and Clustering (two of the most popular DM tasks) and focus on the use of ACO for Classification and Clustering. Moreover, we briefly present related applications and examples and outline possible future trends of this promising collaborative use of techniques.