Study on the Evaluation Function Parameters of the Checkers Game Program on Weka Platform

Li Shuqin , Xue Weiming,Yuan Xiaohua

Based on the machine learning algorithm provided by the Weka platform, this paper performs two studies on the evaluation function parameters of checkers game program. The first one is that, through selecting six numbers, which include the number of the black, the red, the king of black, the king of red, the black in danger, and the red in danger, assigning different weight to each number, then inputting these weighted numbers as factors into the machine learning algorithm, after tested, picking the important numbers as the evaluation parameters, thus, we optimize the parameters of evaluation function for checkers game. The second one is that, by analyzing the influence of removing the weighted numbers of the king of black and the king of red from the program, we obtain the conclusion that the two parameters have large influence, and cannot be removed from the evaluation function.