Improved Decision tree algorithm for data streams with Concept-drift adaptation

Decision tree construction is a well studied problem in data mining. Recently, there has been much interest in mining streaming data. Algorithms like VFDT and CVFDT exist for the construction of a decision tree but, as the new examples are added, a new model has to be generated. In this paper, we have given an algorithm for construction of a decision tree that uses discriminant analysis, to choose the cut point for splitting tests thereby optimizing the time complexity to O(n) from O(nlogn). Also various adaptive learning strategies like contextual, dynamic ensemble, forgetting and detector approaches have been analyzed and handling of concept-drift occurred due to gradual change in data-set is discussed using naïve Bayes classifier at each inner node.