Developing a feature weight self-adjustment mechanism for the CD-sIB algorithm

The sIB algorithm is one of the popular clustering algorithms due to its superior scalability and efficiency. The CD-sIB algorithm is proposed to solve the problem that the sIB algorithm can not handle Non co-occurrence data. It proposes a feature construction method to extend the dataset attributes with a binary transformation. However, the CD-sIB algorithm treats all features evenly and sets weights of all features equally. To address the issue, the paper proposes a feature weight self-adjustment mechanism for the CD-sIB algorithm. A weight-adjusting procedure is applied in the pre-processing stage. In the procedure, the weights of features are adjusted iteratively. The purpose of the feature self-adjusting mechanism is to simultaneously minimize the separations within clusters and maximize the separations between clusters. So that it can improve the quality of the clustering result. Experiments on the Non co-occurrence datasets show that the proposed algorithm based on the feature self-adjusting mechanism is superior to the CD-sIB algorithm.