In this paper, we propose an ant evolutionary classification model, which treats different classes as ant colonies to classify the unlabeled instances. In our model, each ant colony sends its members to propagate its unique pheromone on the unlabeled instances. The unlabeled instances are treated as unlabeled ants. They are assigned to different ant colonies according to the pheromone that different colonies leave on it. Next, the natural selection is carried out to maintain the history colony information as well as the scale of swarms. Theoretical analysis and experimental results show the effectiveness of our proposed model for evolutionary data classification.
CITATION STYLE
He, P., Xu, X. H., Lu, L., Qian, H., Zhang, W., & Li, K. W. (2014). Semi-supervised ant evolutionary classification. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8795). Springer Verlag. https://doi.org/10.1007/978-3-319-11897-0_1
Mendeley helps you to discover research relevant for your work.