Technologies in decision trees construction

Classification is an important problem in data mining. Given a database of records, each with a class label, a classifier generates a concise and meaningful description for each class that can be used to classify subsequent records. Decision trees are a very popular class of classifiers. There are two phases commonly in the procedure of decision trees construction. These classifiers first build a decision tree in a tree-growing phase and then prime subtrees from the decision tree in a subsequent tree-pruning phase to improve accuracy and prevent "over-fitting". There have been a large number of technologies in decision trees construction in the past literature. We present several good algorithms in this paper comparing and analyzing their performance in the construction procedure. Which would be helpful for the future work of decision trees construction.