Data Mining using Pruned Artificial Neural Network Tree (ANNT)
S. Kalaiarasi Anbananthen, G. Sainarayanan, Ali Chekima, Jason Teo · 2006
Artificial Neural Network (ANN) have not been effectively utilized in data mining because of the "black box" nature. This issue was resolved by using Artificial Neural Network Tree (ANNT) approach in the our earlier works. Future improvement was made by incorporating pruning in ANNT approach. ANNT pruning approach consists of three phases: training, pruning and rule extraction. The training phase is concerned with ANN learning followed by pruning. In pruning, the redundant links from the trained network are deleted, Rules are extracted from the pruned network. The proposed scheme results in extracting rules from contributing links and indirectly reduce the number of rules but maintaining classification accuracy.