Ant Based Semi-supervised Classification

Semi-supervised classification methods make use of the large amounts of relatively inexpensive available unlabeled data along with the small amount of labeled data to improve the accuracy of the classification. This article presents a novel 'self-training' based semi-supervised classification algorithm using the property of aggregation pheromone found in natural behavior of real ants. The proposed algorithm is evaluated with real life benchmark data sets in terms of classification accuracy. Also the method is compared with two conventional supervised classification methods and two recent semi-supervised classification techniques. Experimental results show the potentiality of the proposed algorithm.