A Hybrid Approach to Improve Classification with Cascading of Data Mining Tasks
D. Lavanya, K. Usha Rani · 2013
Data mining plays an important role to find the interesting patterns from databases. Medical data mining is very much useful to medical practitioners. To diagnose the patient’s disease Classification, one of the data mining tasks, plays a significant role. Cascading classification with some other data mining tasks improves classification accuracy. In this study a hybrid approach which is a combination of CART decision tree classifier with clustering and feature selection has been proposed on breast cancer data sets. The effectiveness of hybrid approach has been compared against CART with Feature Selection, Classification with Clustering and without Feature Selection in terms of accuracy.