将数据放入 DataFrame 会在 SciKit 学习算法中给出不同的结果

Question

我刚刚注意到 Sci-Kit Learn 的线性回归算法将一些不同的结果加载到 Pandas 数据帧中，而不是仅在原始状态下使用它们。我不明白为什么会这样。考虑以下线性回归示例：from sklearn.datasets import load_bostonfrom sklearn.linear_model import LinearRegressionboston     = load_boston()X1         = pd.DataFrame(boston.data)X1.columns = boston.feature_namesX2         = boston.datay2         = boston.targety1         = boston.targetlreg       = LinearRegression()X1         = (X1 - X1.mean()) / X1.std()X2         = (X2 - X2.mean()) / X2.std()结果模型给出了相同的 R^2 值和预测值，但系数和截距的结果却大不相同。展示：intcpt1 = lreg.fit(X1, y1).intercept_intcpt2 = lreg.fit(X2, y2).intercept_f"Intercept for model with dataframe: {intcpt1}, model with numpy array: {intcpt2}"给出：'Intercept for model with dataframe: 22.53280632411069, model with numpay array: -941.8009906279219'同样，系数也大不相同：coef1 = lreg.fit(X1, y1).coef_[:3]coef2 = lreg.fit(X2, y2).coef_[:3]f"First the coeffs for model with dataframe: {coef1}, modely with numpy array: {coef2}"这使：'First the coeffs for model with dataframe: [-0.92906457  1.08263896  0.14103943], modely with numpy array: [-15.67844685   6.73818665   2.98419849]'但是得分和预测是一样的：score1 = lreg.fit(X1, y1).score(X1, y1)score2 = lreg.fit(X2, y2).score(X2, y2)f"Score for model with dataframe: {score1}, model with numpy array: {score2}"产量：'Score for model with dataframe: 0.7406426641094094, model with numpy array: 0.7406426641094073'同样对于系数：pred1 = lreg.fit(X1, y1).predict(X1)[:3]pred2 = lreg.fit(X2, y2).predict(X2)[:3]f"First 3 predictions with dataframe: {pred1}, with numpy array: {pred2}"提供：'First 3 predictions with dataframe: [30.00384338 25.02556238 30.56759672], with numpy array: [30.00384338 25.02556238 30.56759672]'

Answer 1

这是因为你的转变：

X1 = (X1 - X1.mean()) / X1.std()
X2 = (X2 - X2.mean()) / X2.std()

Pandas 将计算沿列的均值和标准差。要为 numpy 执行此操作，请将轴参数添加到mean和std：

X2 = (X2 - X2.mean(axis=0)) / X2.std(axis=0)