A method for handling clusturing of uncertain data

In this paper, we propose two novel active learning algorithms: 1) k-mode for classifying the certain and uncertain dataset in a whole dataset, 2) Priority R-Tree clustering the certain and uncertain data for each domain. They handle both supervised and unsupervised dataset. These techniques improve the robustness and accuracy of the clustering outcome to a great extent. By minimizing the expected error with respect to the optimal classifier, experimental results display the cluster using the Gas sensor array drift Dataset.