Rapid Clustering with Semi-Supervised Ensemble Density Centers

Clustering algorithm regards robustness, stability, accuracy as performance measurement. In most recent studies, there has been no detailed investigation of resource usage such as CPU, memory, and executing time in most recent studies, one unanticipated finding was that several consensus functions waste many resources to give the outcomes of a small features dataset. In order to resolve these issues, we suggest two new measurement aspects that termed as resource usage and rapidity. Thus, we proposed a new method characterized by rapid implementation, as well as improved the accuracy with the best pure result. This model termed as Rapid Clustering with Semi-supervised Ensemble Density Centers. The strengths of our model is based on fewer iterations and depend on formulas and built-in functions instead of complex coding procedures, then pick objects with its features as cluster centers, and finally, utilize the semi-supervised learning specifically the pairwise constraints, and take advantage of density calculation. Furthermore, the experimental results demonstrated that our model gained the results rapidly with best accuracy and purity.