An alternative clustering algorithm based on IB method

Alternative clustering aims at exploring another reasonable clustering which is distinctively different from an existing one. This paper presents a novel alternative clustering algorithm based on the IB method, named Alt_sIB. Our approach aims to ensure the clustering quality by maximizing the mutual information between clustering labels and data observation, whilst ensuring the clustering distinctiveness by minimizing the information sharing between the two clusterings. We employ a nonparametric MeanNN differential entropy estimator for the mutual information estimation and optimize the objective function iteratively in a sequential way. The experimental results indicate that the proposed Alt_sIB algorithm could uncover the reasonable and different clusterings from the dataset efficiently. Compared to the existing NACI algorithm and minCEntropy algorithm, the Alt_sIB's performance is better.