A Machine Learning Algorithm Based on Supervised Clustering and Classification

In this paper a novel data mining technique - Clustering and Classification Algorithm-Supervised (CCA-S)1 is introduced. CCA-S supports incremental learning and non-hierarchical clustering, and is scalable for processing large data sets. CCA-S incorporates the class information in making clustering decisions, and uses the resulting clusters to classify new data records. We apply and test CCA-S on several common data sets for classification problems. The testing results show that the classification performance of CCAS is comparable to the other classification algorithms such as decision trees, artificial neural networks and discriminant analysis.