earticle

영어

This research has for objective to evaluate the win/loss pattern depending on scores per quarter of each basketball team, which has played ten minutes for four quarters during total two seasons, 2010season, 2011season and, simultaneously, to figure out which quarter’s goal average would be a major factor in winning or losing. It, also, has for objective to set a segmented analysis index for evaluating basketball games. In order to accomplish its objective, this study included data which shows score per quarter of ten teams in total 1080 games for the 2010season, 2011season of the Korean Basketball League, and the decision tree model, one of the Data Mining techniques, through SPSS clementine 11.1 was used to establish the win/loss pattern model depending on scores per quarter and to analyze predictive capability of game results. An algorithm of the decision tree model this research used was CART algorithm. The analysis result is the same as the next. The pattern of which the winning rate is high appeared for 3 pattern. first getting over 23.5 points in 3 quarter. second pattern is getting over 20.5points in 2 quarter. last pattern is getting over 15.5 points in 4 quarter. The pattern of which the defeat rate is low appeared for 4 pattern, too. First getting lower 23.5 points in 3 quarter. Second pattern is getting lower 23.5 points in 1 quarter. Third pattern is getting lower 14.5 points in 2 quarter. last pattern is getting lower 20.5 points in 1 quarter. Consequently, The most important factor of connection with korea professional basket ball victory and defeat is the score distribution of at the 3 quarter. additionally, the major element appear at the distribution of score of 1, 2 quarter.