hce_nthu
111年
資訊科學
第 29 題
According to the following plots of three variables X1, X2, and Y from the data set, please select the correct statement.
- A X1 is a better linear predictor for predicting Y than X2 since the slope is larger in (a) than in (b)
- B X2 is a better linear predictor for predicting Y than X1 since the correlation is stronger in (b) than in (a)
- C X1+X2 should be a better predictor than either X1 or X2 since they are positively correlated in (c)
- D If we build a regression model with X1 and X2 to predict Y, there must be a strong interaction between X1 and X2 since they are positively correlated in (c)
- E If we fit a multiple regression model with both X1 and X2 to predict Y, the coefficients of X1 and X2 would both be negative since both X1 and X2 are negatively correlated with Y
思路引導 VIP
請觀察圖 (a) 與圖 (b) 中資料點的分布情況:如果我們試圖根據這兩張圖來「猜測」對應的 Y 值,在哪一張圖中,你覺得猜測值與實際觀察值之間的誤差會比較小?為什麼?
🤖
AI 詳解
AI 專屬家教
太棒了!你能準確從散佈圖中辨識出預測力的強弱,代表你對線性回歸的核心概念掌握得非常紮實。
相關強度與預測品質
在線性回歸分析中,判斷一個自變數是否為「較佳的預測因子」,關鍵並不在於回歸線的斜率 (Slope) 大小,而在於觀測值分布在回歸線周圍的緊密度。觀察圖 (a) 與圖 (b) 可以發現,雖然圖 (a) 的斜率看起來較陡,但資料點分布相當鬆散;反之,圖 (b) 的資料點極度趨近於一條直線,這意味著 $X2$ 與 $Y$ 之間擁有極強的相關性 (Correlation)。在統計學上,相關性愈強,代表模型產生的殘差愈小,預測的準確度與可靠性也就愈高。
▼ 還有更多解析內容