八种数值变量的特征工程技术：利用Sklearn、Numpy和Python将数值转化为预测模型的有效特征

import matplotlib.pyplot as plt  
 import seaborn as sns  
 from sklearn.datasets import load_wine  
 from sklearn.model_selection import train_test_split  
 from sklearn.preprocessing import StandardScaler  
 from sklearn.neighbors import KNeighborsClassifier  
 from sklearn.metrics import confusion_matrix  
 import numpy as np  
   
 X, y = load_wine(return_X_y=True)  
   
 # 将数据划分为训练集和测试集  
 X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=42)  
   
 # 定义训练模型并获取混淆矩阵的函数  
 def get_confusion_matrix(X_train, X_test, y_train, y_test):  
     model = KNeighborsClassifier(n_neighbors=5)  
     model.fit(X_train, y_train)  
     y_pred = model.predict(X_test)  
     return confusion_matrix(y_test, y_pred)  
   
 # 获取未归一化的混淆矩阵  
 cm_without_norm = get_confusion_matrix(X_train, X_test, y_train, y_test)  
   
 # 归一化数据  
 scaler = StandardScaler()  
 X_train_scaled = scaler.fit_transform(X_train)  
 X_test_scaled = scaler.transform(X_test)  
   
 # 获取归一化后的混淆矩阵  
 cm_with_norm = get_confusion_matrix(X_train_scaled, X_test_scaled, y_train, y_test)  
   
 # 创建两个并列的子图  
 fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(20, 8))  
   
 # 定义绘制热图的函数  
 def plot_heatmap(ax, cm, title):  
     sns.heatmap(cm, annot=True, fmt='d', cmap='viridis', ax=ax, cbar=False)  
     ax.set_title(title, fontsize=16, pad=20)  
     ax.set_xlabel(''Predicted', fontsize=12, labelpad=10)  
     ax.set_ylabel('Actual', fontsize=12, labelpad=10)  
   
 # 绘制热图  
 plot_heatmap(ax1, cm_without_norm, 'Confusion Matrix\nWithout Normalization')  
 plot_heatmap(ax2, cm_with_norm, 'Confusion Matrix\nWith Normalization')  
   
 # 添加共用的颜色条  
 cbar_ax = fig.add_axes([0.92, 0.15, 0.02, 0.7])  
 sm = plt.cm.ScalarMappable(cmap='viridis', norm=plt.Normalize(vmin=0, vmax=np.max([cm_without_norm, cm_with_norm])))  
 fig.colorbar(sm, cax=cbar_ax)  
   
 # 调整布局并显示图表  
 plt.tight_layout(rect=[0, 0, 0.9, 1])  
 plt.show()
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.

from sklearn.preprocessing import PolynomialFeatures  
 import numpy as np  
   
 X = np.array([[1, 2], [3, 4]])  
 poly = PolynomialFeatures(degree=2, include_bias=False)  
 X_poly = poly.fit_transform(X)  
   
 print(X_poly)  
 # 输出: [[1. 2. 1. 2. 4.]  
 #       [3. 4. 9. 12. 16.]]  
   
 print(poly.get_feature_names(['x1', 'x2']))  
 # 输出: ['x1', 'x2', 'x1^2', 'x1 x2', 'x2^2']
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.

import matplotlib.pyplot as plt  
   
 fig, ax = plt.subplots(figsize=(12, 4))  
 average_week_demand = df.groupby(["weekday", "hour"])["count"].mean()  
 average_week_demand.plot(ax=ax)  
 _ = ax.set(  
     title="Average hourly bike demand during the week",  
     xticks=[i * 24 for i in range(7)],  
     xticklabels=["Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat"],  
     xlabel="Time of the week"
 )
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.

import numpy as np  
 import pandas as pd  
   
 from sklearn.preprocessing import FunctionTransformer  
   
 import matplotlib.pyplot as plt  
 
 def sin_transformer(period):  
     return FunctionTransformer(lambda x: np.sin(x / period * 2 * np.pi))  
 
 def cos_transformer(period):  
     return FunctionTransformer(lambda x: np.cos(x / period * 2 * np.pi))  
   
 hour_df = pd.DataFrame(  
     np.arange(26).reshape(-1, 1),  
     columns=["hour"],  
 )  
 hour_df["hour_sin"] = sin_transformer(24).fit_transform(hour_df)["hour"]  
 hour_df["hour_cos"] = cos_transformer(24).fit_transform(hour_df)["hour"]  
 hour_df.plot(x="hour")  
 _ = plt.title("Trigonometric encoding of the 'hour' feature")
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.

import matplotlib.pyplot as plt  
 import numpy as np  
 from sklearn.linear_model import LinearRegression  
 from sklearn.preprocessing import KBinsDiscretizer  
 from sklearn.tree import DecisionTreeRegressor  
   
 # 设置随机数生成器以确保可重复性  
 rnd = np.random.RandomState(42)  
   
 # 创建数据集  
 X = rnd.uniform(-3, 3, size=100)  # 在-3和3之间的100个点  
 y = np.sin(X) + rnd.normal(size=len(X)) / 3  # 正弦函数加噪声  
 X = X.reshape(-1, 1)  # 重塑为sklearn所需的正确格式  
   
 # 应用KBinsDiscretizer  
 discretizer = KBinsDiscretizer(n_bins=10, encode="onehot")  
 X_binned = discretizer.fit_transform(X)  
   
 # 准备可视化  
 fig, (ax1, ax2) = plt.subplots(ncols=2, sharey=True, figsize=(12, 5))  
 line = np.linspace(-3, 3, 1000).reshape(-1, 1)  # 用于绘图的点  
   
 # 训练和绘制模型的函数  
 def train_and_plot(X_train, X_plot, ax, title):  
     # 线性回归  
     linear_reg = LinearRegression().fit(X_train, y)  
     ax.plot(line, linear_reg.predict(X_plot), linewidth=2, color="green", label="Linear Regression")  
       
     # 决策树  
     tree_reg = DecisionTreeRegressor(min_samples_split=3, random_state=0).fit(X_train, y)  
     ax.plot(line, tree_reg.predict(X_plot), linewidth=2, color="red", label="Decision Tree")  
       
     # 原始数据  
     ax.plot(X[:, 0], y, "o", c="k", alpha=0.5)  
     ax.legend(loc="best")  
     ax.set_xlabel("Input Feature")  
     ax.set_title(title)  
   
 # 绘制原始数据的图  
 train_and_plot(X, line, ax1, "Results Before Discretization")  
 ax1.set_ylabel("Regression Output")  
   
 # 绘制离散化数据的图  
 line_binned = discretizer.transform(line)  
 train_and_plot(X_binned, line_binned, ax2, "Results After Discretization")  
   
 plt.tight_layout()  
 plt.show()
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.

import numpy as np  
 import pandas as pd  
 import matplotlib.pyplot as plt  
   
 # 创建一个具有正偏度值的示例数据集  
 np.random.seed(42)  
 data = {  
     'Income': np.random.exponential(scale=50000, size=1000)  # 指数分布以模拟偏度  
 }  
 df = pd.DataFrame(data)  
   
 # 创建一个具有两个并排子图的图形  
 fig, axes = plt.subplots(1, 2, figsize=(15, 5))  
   
 # 原始分布的图  
 axes[0].hist(df['Income'], bins=50, color='blue', alpha=0.7)  
 axes[0].set_title('Original Income Distribution')  
 axes[0].set_xlabel('Income')  
 axes[0].set_ylabel('Frequency')  
   
 # 应用对数变换  
 df['Log_Income'] = np.log1p(df['Income'])  # log1p相当于log(x + 1)  
   
 # 变换后分布的图  
 axes[1].hist(df['Log_Income'], bins=50, color='green', alpha=0.7)  
 axes[1].set_title('Log-transformed Income Distribution')  
 axes[1].set_xlabel('Log_Income')  
 axes[1].set_ylabel('Frequency')  
   
 # 显示图形  
 plt.tight_layout()  
 plt.show()
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.

import numpy as np  
 import matplotlib.pyplot as plt  
 from sklearn.preprocessing import PowerTransformer  
   
 # 生成包含正值和负值的数据  
 np.random.seed(42)  
 data_positive = np.random.exponential(scale=2, size=1000)  # 正值  
 data_negative = -np.random.exponential(scale=0.5, size=200)  # 负值  
 data = np.concatenate([data_positive, data_negative])  # 合并正值和负值数据  
   
 # 创建两个PowerTransformer实例：一个用于Yeo-Johnson，一个用于Box-Cox进行比较  
 pt_yj = PowerTransformer(method='yeo-johnson', standardize=False)  # Yeo-Johnson变换  
 pt_bc = PowerTransformer(method='box-cox', standardize=False)  # Box-Cox变换  
   
 # 应用变换  
 data_yj = pt_yj.fit_transform(data.reshape(-1, 1))  # 应用Yeo-Johnson变换  
   
 # Box-Cox需要正数据，所以我们添加一个偏移量使所有值为正  
 data_offset = data - np.min(data) + 1e-6  # 偏移量以确保所有值为正  
 data_bc = pt_bc.fit_transform(data_offset.reshape(-1, 1))  # 应用Box-Cox变换  
   
 # 可视化结果  
 fig, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize=(18, 5))  
   
 # 原始数据的直方图  
 ax1.hist(data, bins=50, edgecolor='black')  
 ax1.set_title("Original Data")  
 ax1.set_xlabel("Value")  
 ax1.set_ylabel("Frequency")  
   
 # Yeo-Johnson变换后数据的直方图  
 ax2.histax2.hist(data_yj, bins=50, edgecolor='black')  
 ax2.set_title("Yeo-Johnson")  
 ax2.set_xlabel("Value")  
   
 # Box-Cox变换后数据的直方图  
 ax3.hist(data_bc, bins=50, edgecolor='black')  
 ax3.set_title("Box-Cox")  
 ax3.set_xlabel("Value")  
   
 plt.tight_layout()  
 plt.show()
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.

import numpy as np  
 from sklearn.preprocessing import QuantileTransformer  
 import matplotlib.pyplot as plt  
   
 # 创建一个具有偏斜分布的样本数据集  
 np.random.seed(0)  
 data = np.random.exponential(scale=2, size=(1000, 1))  # 指数分布  
   
 # 初始化QuantileTransformer  
 quantile_transformer = QuantileTransformer(n_quantiles=100, output_distribution='normal')  
   
 # 应用变换  
 data_transformed = quantile_transformer.fit_transform(data)  
   
 # 可视化原始和变换后的数据  
 plt.figure(figsize=(12, 6))  
   
 # 原始数据的直方图  
 plt.subplot(1, 2, 1)  
 plt.hist(data, bins=50, color='blue', edgecolor='black')  
 plt.title("Original Data (Exponential)")  
   
 # 变换后数据的直方图  
 plt.subplot(1, 2, 2)  
 plt.hist(data_transformed, bins=50, color='green', edgecolor='black')  
 plt.title("Transformed Data (Normal)")  
   
 plt.show()
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.

import numpy as np  
 import matplotlib.pyplot as plt  
 from sklearn.datasets import fetch_lfw_people  
 from sklearn.decomposition import TruncatedSVD  
   
 # 加载Labeled Faces in the Wild数据集  
 lfw_people = fetch_lfw_people(min_faces_per_person=70, resize=0.4)  
 X = lfw_people.data  
 n_samples, h, w = lfw_people.images.shape  
   
 print(f"数据集维度：{X.shape}")  
   
 # 用于降维的组件数  
 n_components = 50  
   
 # 应用TruncatedSVD  
 svd = TruncatedSVD(n_components=n_components, random_state=42)  
 X_transformed = svd.fit_transform(X)  
   
 # 重构图像的函数  
 def reconstruct_image(svd, X_transformed):  
     return np.dot(X_transformed, svd.components_)  
   
 # 重构所有图像  
 X_reconstructed = reconstruct_image(svd, X_transformed)  
   
 # 显示原始和降维人脸的函数  
 def plot_transposed_original_and_reduced(X_original, X_reduced, n_faces=5):  
     fig, axes = plt.subplots(2, n_faces, figsize=(3*n_faces, 6))  
     random_indices = np.random.choice(X_original.shape[0], n_faces, replace=False)  
       
     for i, idx in enumerate(random_indices):  
         # 原始人脸  
         axes[0, i].imshow(X_original[idx].reshape(h, w), cmap='gray')  
         axes[0, i].set_title(f'Face {i+1}\nOriginal')  
         axes[0, i].axis('off')  
           
         # 降维人脸  
         axes[1, i].imshow(X_reduced[idx].reshape(h, w), cmap='gray')  
         axes[1, i].set_title(f'Reduced\n({n_components} comp.)')  
         axes[1, i].axis('off')  
       
     plt.tight_layout()  
     plt.show()  
   
 # 显示原始和降维人脸  
 plot_transposed_original_and_reduced(X, X_reconstructed)  
   
 # 计算并打印平均重构误差  
 mse = np.mean((X - X_reconstructed) ** 2)  
 print(f"\n平均重构误差 (MSE)：{mse:.4f}")  
   
 # 打印累积解释方差  
 explained_variance_ratio = svd.explained_variance_ratio_  
 cumulative_variance_ratio = np.cumsum(explained_variance_ratio)  
 print(f"\n{n_components} 个组件的累积解释方差：{cumulative_variance_ratio[-1]:.4f}")  
   
 # 显示前几个组件  
 n_components_show = min(10, n_components)  
 components = svd.components_[:n_components_show].reshape((n_components_show, h, w))  
 fig, axes = plt.subplots(2, 5, figsize=(15, 6))  
 for i, ax in enumerate(axes.flat):  
     if i < n_components_show:  
         ax.imshow(components[i], cmap='gray')  
         ax.set_title(f'Component {i+1}')  
     ax.axis('off')  
 plt.tight_layout()  
 plt.show()
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.