한국프로야구에서 승패 추정에 관한 연구

A Study on Winning and Losing in Korean Professional Baseball League

2000년부터 2005년까지 한국프로야구 8개팀의 6328게임 중에 무승부 182게임을 제외한 6146게임의 환경부분(홈 혹은 원정), 공격부분(안타, 도루, 희생번트, 희생플라이, 4구, 사구, 삼진, 병살), 수비부분(실책), 투수부분(피안타, 피도루, 피4구, 피사구, 탈삼진)의 자료를 이용하여 경기 승패에 영향을 주는 변수와 순위를 알고자 로지스틱회귀분석, 판별분석, 의사결정나무분석 방법을 실시하여 그 결과를 비교 분석하였다.

Total games of 6328 of eight teams of Korean Pro Baseball from season of 2000 to 2005, excepting 182 games ended in a tie, this study analyzed 6,146 games in the field of environment (home or road game), attack (hit, stolen base, sacrifice bunt, sacrifice fly, four balls, hit by pitch, strikeout, double play), defense (error), pitcher (being hit, being stolen base, being four balls, being hit by pitch, being strikeout), in order to find out the variables affecting the result of a game and their ranking, this study conducted logistic regression, discriminant analysis, and decision tree, to compare and analyze the result.