A Refinement Algorithm of Decision Tree Generation
Fangxia Shi · Journal of Xianyang Normal University · 2006
Decision tree induce algorithm is the key algorithm of decision tree,which is widely applicable.Its main deficiency is that the algorithm is based on PM(Probabilistic Model) and local optimization.This paper introduce a method of average information gain to improve the decision tree induce algorithm.The standard of the test attribute is the max value of the average information gain.The experiment proved that algorithm has more value in the experiment.