Constrained ant colony optimization for data clustering

We use the heuristics known as ant colony optimization in the partitioning problem for improving solutions of k-means method. Each ant in the algorithm is associated to a partition, which is modified by the principles of the heuristics; that is, by the random selection of an element, and the assignment of another element which is chosen according to a probability that depends on the pheromone trail (related to the overall criterion: the maximization of the between-classes variance), and a local criterion (the distance between objects). The pheromone trail is reinforced for those objects that belong to the same class. We present some preliminary results, compared to results of other techniques, such as simulated annealing, genetic algorithm, tabu search and k-means. Results are as good as the best of the above methods.

The clustering algorithms have evolved over the last decade. With the continuous success of natural inspired algorithms in solving many engineering problems, it is imperative to scrutinize the success of these methods applied to data clustering. These naturally inspired algorithms are mainly stochastic search and optimization techniques, guided by the principles of collective behavior and self-organization of insect swarms. The parameters setting of the ant colony clustering algorithms determine the behavior of each ant and are critical for fast convergence to near optimal solutions of clustering task. This inspired us to explore techniques for automatically learning the optimal parameters for a given clustering task. We devised and implemented a hybrid Ant-Colony clustering algorithm, which uses particle swarm optimization algorithm in the early stages to 'breed' a population of ants possessing near optimal behavioral parameter settings for a given problem. This hybrid algorithm converges rapidly for nearly optimal parameters that maximize the ant-colony clustering behavior.

Advanced Studies in Behaviormetrics and Data Science, 2020

An ant colony optimization approach for partitioning a set of objects is proposed. In order to minimize the intra-variance, or within sum-of-squares, of the partitioned classes, we construct ant-like solutions by a constructive approach that selects objects to be put in a class with a probability that depends on the distance between the object and the centroid of the class (visibility) and the pheromone trail; the latter depends on the class memberships that have been defined along the iterations. The procedure is improved with the application of K-means algorithm in some iterations of the ant colony method. We performed a simulation study in order to evaluate the method with a Monte Carlo experiment that controls some sensitive parameters of the clustering problem. After some tuning of the parameters, the method has also been applied to some benchmark real-data sets. Encouraging results were obtained in nearly all cases.

International Journal of Computer Applications, 2012

Grouping different objects possessing inherent similarities in clusters has been addressed as the clustering problem among researchers. The development of new metaheuristics has given another direction to data clustering research. Swarm intelligence technique using ant colony optimization provides clustering solutions based on brood sorting. After basic ant model of clustering, number of improvements has been proposed. But the ant clustering still suffers with low convergence. This paper presents a novel model of intelligent movement of ants including the negative pheromone and direction selection. Negative pheromone plays a role of barrier in the direction of empty area and direction selection avoids the calculations not contributing to the clustering process. Simulations have shown good results.

2010

The existing Ant Colony Optimization (ACO) Algorithms for the Quadratic Assignment Problem (QAP) are often combined with two kinds of Stochastic Local Search (SLS) methods: the 2-opt local search and the tabu local search. In this paper, these two SLS methods are respectively improved according to the properties of ACO and QAP. For the 2-opt local search, a new random walk strategy is used to avoid a quick stagnation into local optima. Moreover, a forwardlooking strategy is proposed to explore the neighborhood more thoroughly. In the case of tabu local search, a random walk strategy is also employed to avoid getting stuck at local optima. Experimental evaluation of the ACO algorithms combined with the improved local search proposed in this paper are conducted on problems from the well known QAPLIB library. The results demonstrate that each ACO algorithm, combined with its respective improved local search, has a better performance in terms of the quality of the solution returned than the ACO algorithm with the original local search techniques. Moreover, we also noticed that the improved methods outperform each other for different classes of problems.

2011

This paper presents a new algorithm for clustering which is called an “efficient ant colony optimization clustering algorithm” (EACOC) based on a classic algorithm “LF algorithm”. We have proved the algorithm efficiency when dealt with a big variety of different data as well as providing high quality and converging speed simultaneously. This is considered as the outcome of many changes we have made including redefining the digital manner of ants, setting new formula to find out the degree of similarity and measuring the distance between objects; as well as creating a process to determine the degree of similarity between the collections resulting from the repeated processes. Experimental results show, by using clustering benchmarks indicate, that this suggested algorithm is the best of (LF) Algorithm, as it could defeat the defects found in (LF) involving; the law converging speed and the big number of repeating processes.

2011

Cluster Analysis is a popular data analysis and data mining technique. High quality and fast clustering algorithms play a vital role for users to navigate, effectively organize the data and summarize data. Ant Colony Optimization (ACO), a Swarm Intelligence technique, integrated with clustering algorithms, is being used by many applications for past few years. In this paper we discuss recent improvements on clustering algorithms like PP (Project Pursuit) based on the ACO algorithm for high dimensional data, recent applications of Data Clustering with ACO, application of Ant-based clustering algorithm for object finding by multiple robots in image processing field and the hybrid PSO/ACO algorithm for better optimized results.

2010

Among the many bio–inspired techniques, ant clustering algorithms have received special attention, especially because they still require much investigation to improve performance, stability and other key features that would make such algorithms mature tools for data mining. Clustering with swarm–based algorithms is emerging as an alternative to more conventional clustering methods, such as k–means algorithm. This proposed approach mimics the clustering behavior observed in real ant colonies. As a case study, this paper focuses on the behavior of clustering procedures in this new approach. The proposed algorithm is evaluated on a number of well–known benchmark data sets. Empirical results clearly show that the ant clustering algorithm (ACA) performs well when compared to other techniques.

Lecture Notes in Computer Science, 2005

In this paper the novel concept of ACO and its learning mechanism is integrated with the K-means algorithm to solve image clustering problems. The learning mechanism of the proposed algorithm is obtained by using the defined parameter called pheromone, by which undesired solutions of the K-means algorithm is omitted. The proposed method improves the K-means algorithm by making it less dependent on the initial parameters such as randomly chosen initial cluster centers, hence more stable.

2009 IEEE International Conference on Systems, Man and Cybernetics, 2009

Data clustering plays an important role in many disciplines, including data mining, machine learning, bioinformatics, pattern recognition, and other fields. When there is a need to learn the inherent grouping structure of data in an unsupervised manner, ant-based clustering stand out as the most widely used group of swarm-based clustering algorithms. Under this perspective, this paper presents a new Adaptive Ant-based Clustering Algorithm (AACA) for clustering data sets. The algorithm takes into account the properties of aggregation pheromone and perception of the environment together with other modifications to the standard parameters that improves its convergence. The performance of AACA is studied and compared to other methods using various patterns and data sets. It is also compared to standard clustering using a set of analytical evaluation functions and a range of synthetic and real data collection. Experimental results have shown that the proposed modifications improve the performance of ant-colony clustering algorithm in term of quality and run time.