Revisiting Decision Trees for Large Set of Keys

Recently, we have proposed a method for representing a large set of keys consisting of associating a finitestate automaton accepting the input language with a decision tree representing the output language. However, the number of the leaves of that decision tree is more than the number of distinct values of the output language. This paper proposes simple data structure based on m-ary decision tree and information (corresponding to keys) array for accepting required values. The experiments done on several datasets confirm the effectiveness of the present work.