The Research On Optimizing Based On Program Transformation Of The Decision Tree Algorithms

Program transformation is a method by which a program is transferred into another one according to the certain rules. The syntax of the transferred program is equavolent to the original program .The algorithms and data structures are simplified through the use of transformation scheme which can guarantee the accuracy.Finally,the source program is derivated into a procedure_oriented efficient version.The field in optimization of decision tree is focused by the scientists at home and abroad all the time. Scholars bring up a lot of idea to improve the efficiency in accordance with the excessive matching and scale of decision tree.Through the efforts ,they obtained a large number of achievements.In order to obtain a more efficient decision tree, in the field of optimization of decision tree, the main body of a writing use the program transformation method to optimize the algorithm:(1)We start by writing clean and correct (but probably inefficient)programs in homomorphisms recursion structure.(2)Based on the analysis of the program,we choose proper methods to transform it to more efficient equivalents.(3)Apply tupling technique in data structure definition.Obtaining the new efficient program ,we need scan the data set only one times.(4)Through summarizing the job above ,we draw a conclusion to collect the methods by which the complex program could be well transformed.