A two-stage semi-supervised clustering method based on hybrid particle swarm optimization
Jinxin Dong, Minyong Qi, Fengrui Wang · 2017
In real world applications, there are a large number of unlabeled data, but the number of labeled data is relatively small. It is a fact that the labeled data are often difficult to be gained, and the labeling work is often time consuming. So we must use the few given labeled data more effectively in data analysis. Traditional clustering algorithm can group unlabeled data, but the label of each cluster is uncertain. Clustering only gives partitions to dataset. In this paper, we propose a two-stage semi-supervised clustering algorithm. It can take advantage of clustering algorithm firstly. Then guided by the labeled data, it can fine clustering results based on hybrid particle swarm optimization. Experiment results show that the proposed method can give good results using both labeled data and unlabeled data.