本文提出多種轉換之線性迴歸模式，分析歷屆奧運男、女田徑及游泳金牌成績(1896 2000)，得到含有最大R2值的初步預測模式，再以獨特的比例方法修正此初步預測模式，而得到最後修正預測模式。本文用最後修正預測模式來預測2004年雅典奧運男、女田徑及游泳的金牌成績。
We present several transformed linear regression models to analyze woman and men's track/field and swimming scores of the Olympic gold medals (1896 to 2000). The model with the largest coefficient of determination (R2), which measures the contribution of independent variables (the years) in predicting the response (the scores); has been used as the preliminary prediction model. Then the proportions of the preliminary prediction scores and the real scores are adopted into our final modified prediction method. We use the final modified transformed linear regression models to predict gold medal scores for the 2004 Olympics.
