New hybrid adaptive Ant Colony Optimizaion and Self-Organizing Map for DNA microarray group finding

The Ant Colony Optimization (ACO) and Neural Networks have been successfully applied to several types of problem such as the difficult NP-hard problem, the optimization problem, and the knowledge discovery problem. This paper proposes the efficient hybrid improved ACO and Self-Organizing Map Neural Network (SOM) to solve the clustering problem. The advantages of this hybrid algorithm are to reduce the disadvantage of ACO and SOM and to provide high accuracy and robustness of cancer predictions. The effectiveness of this hybrid algorithm is illustrated through the results of some DNA microarray datasets and some well-known datasets such as Leukemia, Conlon Cancer, and Iris. The experimental results show that this hybrid algorithm provides high performance in clustering problems.