한국프로농구는 인터넷 사이트(www.kbl.or.kr)를 통해 출전선수별로 경기기록을 자세하게 공개하고 있다. 공개되는 경기기록은 14개의 변인으로 7개의 공격변인과 7개의 수비변인으로 구분할 수 있다. 공격변인으로는 2P%(2점슛성공률), 3P%(3점슛성공률), FT%(자유투성공률), DK(덩크슛), OR(공격리바운드), A(어시스트), TO(턴오버)가 기록되고, 수비변인으로는 DR(수비리바운드), ST(스틸), GD(굿디팬스), BS(블록슛), WFT(자유투 유 파울), WOFT(자유투 무 파울), TF(테크니컬파울)이 기록된다. 본 연구는 이와 같이 공개되는 경기기록을 승리한 집단과 패배한 집단으로 구분하여 통계적인 방법을 통해 승패결정요인을 분석하였다. 통계적인 방법으로 SPSS 15.0 프로그램을 통해 로지스틱 회귀분석을 실시하였고, Answer Tree 3.0 프로그램을 통해 의사결정나무 분석을 실시하였다. 본 연구의 결과를 요약하면 다음과 같다. 로지스틱 회귀분석, 전진적 선정(Forward Selection)방법은 방정식에 포함된 변인(Variables in the Equation)결과에서 설명력이 가장 높은 변인 순으로 결과를 단계별로 제시해 준다. Wald 검증과 우도비 검증(LR test: Likelihood Ratio test)결과 DR, 3P%, ST, 2P%, TO, OR, WFT, FT%, GD순으로 나타났다. 의사결정나무 분석 CHAID 알고리즘 결과, 가장 승률이 높은 모형은 DR가 26개 이상이며, 3P%이 35% 이상일때 경기기록으로, 이길 확률이 89.05%로 나타났다. 가장 승률이 낮은 모형은 DR가 21개이하, 3P%이 44%이하 그리고 ST 횟수가 9회 이하인 경기기록으로, 이길 확률이 17.48%로 나타났다.
한국프로농구는 인터넷 사이트(www.kbl.or.kr)를 통해 출전선수별로 경기기록을 자세하게 공개하고 있다. 공개되는 경기기록은 14개의 변인으로 7개의 공격변인과 7개의 수비변인으로 구분할 수 있다. 공격변인으로는 2P%(2점슛성공률), 3P%(3점슛성공률), FT%(자유투성공률), DK(덩크슛), OR(공격리바운드), A(어시스트), TO(턴오버)가 기록되고, 수비변인으로는 DR(수비리바운드), ST(스틸), GD(굿디팬스), BS(블록슛), WFT(자유투 유 파울), WOFT(자유투 무 파울), TF(테크니컬파울)이 기록된다. 본 연구는 이와 같이 공개되는 경기기록을 승리한 집단과 패배한 집단으로 구분하여 통계적인 방법을 통해 승패결정요인을 분석하였다. 통계적인 방법으로 SPSS 15.0 프로그램을 통해 로지스틱 회귀분석을 실시하였고, Answer Tree 3.0 프로그램을 통해 의사결정나무 분석을 실시하였다. 본 연구의 결과를 요약하면 다음과 같다. 로지스틱 회귀분석, 전진적 선정(Forward Selection)방법은 방정식에 포함된 변인(Variables in the Equation)결과에서 설명력이 가장 높은 변인 순으로 결과를 단계별로 제시해 준다. Wald 검증과 우도비 검증(LR test: Likelihood Ratio test)결과 DR, 3P%, ST, 2P%, TO, OR, WFT, FT%, GD순으로 나타났다. 의사결정나무 분석 CHAID 알고리즘 결과, 가장 승률이 높은 모형은 DR가 26개 이상이며, 3P%이 35% 이상일때 경기기록으로, 이길 확률이 89.05%로 나타났다. 가장 승률이 낮은 모형은 DR가 21개이하, 3P%이 44%이하 그리고 ST 횟수가 9회 이하인 경기기록으로, 이길 확률이 17.48%로 나타났다.
The purpose of this study is to analyze the factor of professional basketball record which is for predicting whether win or lose the game. The factor model of victory and defeat made by collecting present basketball records and it will be used for team strategy. Korea Basketball League(KBL) informs ...
The purpose of this study is to analyze the factor of professional basketball record which is for predicting whether win or lose the game. The factor model of victory and defeat made by collecting present basketball records and it will be used for team strategy. Korea Basketball League(KBL) informs records of every pro-basketball game data through the Internet site(www.kbl.or.kr). The records are 2Point shot percentage(2P%), 3Point shot percentage(3P%), Free throw shot percentage(FT%), and frequency of Dunk shot(DK), Offense rebound(OR), Assist(A), Turn over(TO), Defense rebound(DR), Steal(ST), Good defense(GD), Block shot(BS), With free throw(WFT), Without free throw(WOFT), Technical foul(TF). These 14 variables compose that the 7 formers are offense variables and the 7 latters are defense variables. This study analyzed the records, 10 teams` 270 games data of KBL 2006-2007season and developed the factor model of victory and defeat. The statistic methods used for the prediction were SPSS program for analyzing logistic regression and Answer Tree 3.0 program for analyzing decision tree. Major results are as follow. The factor model of victory and defeat through the logistic regression is dissected in SPSS 15.0 program, Forward Selection method. The result materialized the 9 variables as rank, DR, 3P%, 2P%, ST, TO, OR, FT%, WFT, GD. In analyzing decision tree in Answer Tree 3.0 program, DR, 3P%, 2P%, ST variables are adopted. The highest the percentage of victories in this study is 89.06% that DR is more than 26 times and 3P% is more than 36 percentile. The next high the percentage of victories is 72.22% that DR is 21~25 times and 2P% is more than 53 percentile. In the Pro-Basketball, defense rebound and shot success percentage effect the high percentage of victories. That DR is less than 21 times and 3P% is less than 44% shows 25% of the percentage of victories. In this study, the lowest the percentage of victories is 17.48% that DR is less than 21 times and 3P% is less than 44% and ST is less than 9 times.
The purpose of this study is to analyze the factor of professional basketball record which is for predicting whether win or lose the game. The factor model of victory and defeat made by collecting present basketball records and it will be used for team strategy. Korea Basketball League(KBL) informs records of every pro-basketball game data through the Internet site(www.kbl.or.kr). The records are 2Point shot percentage(2P%), 3Point shot percentage(3P%), Free throw shot percentage(FT%), and frequency of Dunk shot(DK), Offense rebound(OR), Assist(A), Turn over(TO), Defense rebound(DR), Steal(ST), Good defense(GD), Block shot(BS), With free throw(WFT), Without free throw(WOFT), Technical foul(TF). These 14 variables compose that the 7 formers are offense variables and the 7 latters are defense variables. This study analyzed the records, 10 teams` 270 games data of KBL 2006-2007season and developed the factor model of victory and defeat. The statistic methods used for the prediction were SPSS program for analyzing logistic regression and Answer Tree 3.0 program for analyzing decision tree. Major results are as follow. The factor model of victory and defeat through the logistic regression is dissected in SPSS 15.0 program, Forward Selection method. The result materialized the 9 variables as rank, DR, 3P%, 2P%, ST, TO, OR, FT%, WFT, GD. In analyzing decision tree in Answer Tree 3.0 program, DR, 3P%, 2P%, ST variables are adopted. The highest the percentage of victories in this study is 89.06% that DR is more than 26 times and 3P% is more than 36 percentile. The next high the percentage of victories is 72.22% that DR is 21~25 times and 2P% is more than 53 percentile. In the Pro-Basketball, defense rebound and shot success percentage effect the high percentage of victories. That DR is less than 21 times and 3P% is less than 44% shows 25% of the percentage of victories. In this study, the lowest the percentage of victories is 17.48% that DR is less than 21 times and 3P% is less than 44% and ST is less than 9 times.
AI-Helper
※ AI-Helper는 부적절한 답변을 할 수 있습니다.