2002世界盃足球賽比賽結果預測之研究

A Study of Prediction by the Results of Logistic Regression on the Game in 2002 Fifa World Cup

許伯陽(Po-Yang Hsu) ； 周靈山(Lin-Sang Chou)

《大專體育學刊》 4卷2期 (2002/12) Pp. 79-88

https://doi.org/10.5297/ser.200212_4(2).0007

摘要

本研究的目的是以2002年世界盃足球賽13項攻防數據為預測變項，包括進球數、射門數、射在球門內次數、犯規、角球、任意球、點球、越位、烏龍球、黃牌、紅牌、控球百分比以及實際比賽時間等，由多元邏輯迴歸來建立足球賽比賽結果（勝、和、負）的預測模式。邏輯迴歸模式為：
勝與不勝邏輯迴歸方程式：方程式略。
負與不負邏輯迴歸方程式：方程式略。
當進球數大於1.4578時，迴歸模式預測為勝的機率大於不勝的機率；進球數小於0.7364 時，迴歸模式預測為負的機率大於不負的機率；若是在兩者之間，則迴歸模式預測為和的機率大於勝或負的機率。評估預測模式品質的一致性百分比為71.2，不一致性百分比為7.8，同值百分比為 21，其四種指標分別為Somers’D : 0.634, Gamma:0.802，τ-a:0.425，c:0.817。整體而言，2002世界盃足球賽比賽結果的預測模式品質良好。

關鍵字

世界盃足球賽 ； 多元邏輯迴歸

並列摘要

The purpose of the study was to construct a prediction model by the results of the game in 2002 world cup. The dependence variables were the group of win, draw and losing. The independence variables were including 13 offense-defense indexes (goal scored, shot, shot on goal, foul, corner kick, free kick, penalty kick, offside, own goal, caution, expulsion, ball possession %, actual playing time). Model construction could be analyzed by multiple logistic regression. The results of logistic regression equation were
Win and draw or losing: The equation is abbreviated.
Losing and win or draw: The equation is abbreviated.
If the goals scored were more then 1.4578, the probability of win was more then draw or losing in regression prediction model. If the goals scored were less then 0.7364, the probability of losing was more then win or draw in regression prediction model. If the goals scored were between 1.4578 and 0.7364, the probability of draw was more then win or losing in regression prediction model. The quality of prediction model were percent concordant 71.2, percent discordant 7.8 and percent tied 21. The indexes of prediction model were Somers’D 0.634, Gamma 0.802 τ-a 0.425, c 0.8 17. On the whole, the results of the game had good characteristics in 2002 world cup.

並列關鍵字

world cup ； multiple logistic regression

參考文獻

Agresti, A.(1996).An introduction to categorical data analysis..

Hosmer, D. W.,Lemeshow, S.(1989).Applied logistic regression..

Hosmer, T. A.,Hosmer, D. W.,Fisher, L.(1983).A comparison of three methods of estimating the logistic regression coefficient..Communication Statistics Computation.12

Press, S. J.,Wilson, S.(1978).Choosing between logistic regression and discriminate analysis..Journal of the American Statistical Association.73

SAS Institute Inc.(1995).Logistic regression examples using the SAS system..