第 6 章 ソースコード一覧¶

import numpy as np
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split

# データの作成（勉強時間と試験の点数）
np.random.seed(42)
X = np.arange(1, 21).reshape(-1, 1).astype(float)
# 傾き3.5、切片30の直線にノイズを加えてyを作成
y = 3.5 * X.ravel() + 30 + np.random.randn(20) * 5

# データの分割
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42
)

# モデルの学習
model = LinearRegression()
model.fit(X_train, y_train)

# パラメータの確認
print(f"傾き (a): {model.coef_[0]:.2f}")
print(f"切片 (b): {model.intercept_:.2f}")
print(f"回帰式: y = {model.coef_[0]:.2f}x + {model.intercept_:.2f}")

# テストデータでの予測
y_pred = model.predict(X_test)
for xi, yt, yp in zip(X_test.ravel(), y_test, y_pred):
    print(f"  勉強時間 {xi:.0f}h → 実際: {yt:.1f}, 予測: {yp:.1f}")

import numpy as np
import matplotlib.pyplot as plt

# 可視化
plt.figure(figsize=(5.8, 3.6))
plt.scatter(X_train, y_train, color="blue", label="訓練データ")
plt.scatter(X_test, y_test, color="green", marker="s",
            label="テストデータ")

# 回帰直線
X_line = np.linspace(0, 22, 100).reshape(-1, 1)
y_line = model.predict(X_line)
plt.plot(X_line, y_line, color="red", linewidth=2,
         label="回帰直線")

plt.xlabel("勉強時間")
plt.ylabel("試験の点数")
plt.title("単回帰")
plt.legend()
plt.grid(True, alpha=0.3)
plt.show()

import pandas as pd
import numpy as np

# 住宅価格データの作成
np.random.seed(42)
n = 50
data = pd.DataFrame({
    "面積": np.random.randint(40, 130, n),
    "築年数": np.random.randint(0, 35, n),
    "駅徒歩": np.random.randint(1, 25, n)
})
data["価格"] = (
    35 * data["面積"]
    - 30 * data["築年数"]
    - 15 * data["駅徒歩"]
    + 500
    + np.random.randn(n) * 100
)

print(data.head())
print(f"\nデータ数: {len(data)}")

import pandas as pd
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split

# 特徴量と目的変数の分離
X = data[["面積", "築年数", "駅徒歩"]]
y = data["価格"]

# データの分割
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42
)

# モデルの学習
model = LinearRegression()
model.fit(X_train, y_train)

# 係数の確認
coef_df = pd.DataFrame({
    "特徴量": X.columns,
    "係数": model.coef_
})
print(coef_df)
print(f"\n切片: {model.intercept_:.2f}")
print(f"訓練データ R^2: {model.score(X_train, y_train):.4f}")
print(f"テストデータ R^2: {model.score(X_test, y_test):.4f}")

import matplotlib.pyplot as plt

y_pred = model.predict(X_test)

plt.figure(figsize=(2.4, 2.4))
plt.scatter(y_test, y_pred, alpha=0.7)
plt.plot([y_test.min(), y_test.max()],
         [y_test.min(), y_test.max()],
         "r--", linewidth=2, label="y = x")
plt.xlabel("実際の価格")
plt.ylabel("予測価格")
plt.title("実際の値 vs 予測値")
plt.legend()
plt.grid(True, alpha=0.3)
plt.show()

import numpy as np
from sklearn.preprocessing import PolynomialFeatures

X = np.array([[2], [3], [4]])

# 2次の多項式特徴量を生成
poly = PolynomialFeatures(degree=2,
    include_bias=False)
X_poly = poly.fit_transform(X)

print("元の特徴量:")
print(X)
print("\n多項式特徴量 (x, x^2):")
print(X_poly)
print(f"\n特徴量名: {poly.get_feature_names_out()}")

import numpy as np
import matplotlib.pyplot as plt
from sklearn.linear_model import LinearRegression
from sklearn.preprocessing import PolynomialFeatures
from sklearn.metrics import mean_squared_error

# 非線形データの生成
np.random.seed(42)
X = np.sort(np.random.rand(30, 1) * 2 * np.pi, axis=0)
y = np.sin(X).ravel() + np.random.randn(30) * 0.2

# 次数ごとにモデルを学習
degrees = [1, 3, 5, 15]
fig, axes = plt.subplots(2, 2, figsize=(5.8, 4.6))

X_plot = np.linspace(0, 2 * np.pi, 200).reshape(-1, 1)

for ax, degree in zip(axes.ravel(), degrees):
    poly = PolynomialFeatures(degree=degree)
    X_poly = poly.fit_transform(X)
    X_plot_poly = poly.transform(X_plot)

    model = LinearRegression()
    model.fit(X_poly, y)

    y_plot = model.predict(X_plot_poly)
    rmse = np.sqrt(mean_squared_error(y, model.predict(X_poly)))

    ax.scatter(X, y, color="blue", alpha=0.6, s=20)
    ax.plot(X_plot, y_plot, color="red", linewidth=2)
    ax.set_title(f"次数 {degree} (RMSE={rmse:.4f})")
    ax.set_ylim(-2, 2)

plt.tight_layout()
plt.show()

import numpy as np
from sklearn.linear_model import Ridge, LinearRegression
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import PolynomialFeatures, StandardScaler
from sklearn.metrics import mean_squared_error

# 非線形データの生成
np.random.seed(42)
X = np.sort(np.random.rand(30, 1) * 10, axis=0)
y = np.sin(X).ravel() + np.random.randn(30) * 0.3

# 高次の多項式特徴量を生成（過学習しやすい状況を作る）
poly = PolynomialFeatures(degree=10, include_bias=False)
X_poly = poly.fit_transform(X)

# データの分割
X_train, X_test, y_train, y_test = train_test_split(
    X_poly, y, test_size=0.3, random_state=42
)

# 訓練データを基準に特徴量を標準化
scaler = StandardScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)

# 通常の線形回帰
lr = LinearRegression()
lr.fit(X_train, y_train)
print("【通常の線形回帰】")
print(f"訓練 RMSE: "
      f"{np.sqrt(mean_squared_error(y_train, lr.predict(X_train))):.4f}")
print(f"テスト RMSE: "
      f"{np.sqrt(mean_squared_error(y_test, lr.predict(X_test))):.4f}")

# リッジ回帰
ridge = Ridge(alpha=1.0)
ridge.fit(X_train, y_train)
print("\n【リッジ回帰 (alpha=1.0)】")
print(f"訓練 RMSE: "
      f"{np.sqrt(mean_squared_error(y_train, ridge.predict(X_train))):.4f}")
print(f"テスト RMSE: "
      f"{np.sqrt(mean_squared_error(y_test, ridge.predict(X_test))):.4f}")

import numpy as np
from sklearn.linear_model import Lasso
from sklearn.metrics import mean_squared_error

# ラッソ回帰
lasso = Lasso(alpha=0.1, max_iter=100000)
lasso.fit(X_train, y_train)
print("【ラッソ回帰 (alpha=0.1)】")
print(f"訓練 RMSE: "
      f"{np.sqrt(mean_squared_error(y_train, lasso.predict(X_train))):.4f}")
print(f"テスト RMSE: "
      f"{np.sqrt(mean_squared_error(y_test, lasso.predict(X_test))):.4f}")

# 係数の確認（0になった係数の数）
n_zero = np.sum(lasso.coef_ == 0)
print(f"\n全係数の数: {len(lasso.coef_)}")
print(f"0になった係数の数: {n_zero}")
print(f"0でない係数の数: {len(lasso.coef_) - n_zero}")

import numpy as np
import pandas as pd
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_squared_error
from sklearn.metrics import mean_absolute_error, r2_score

# データの準備
np.random.seed(42)
n = 80
data = pd.DataFrame({
    "面積": np.random.randint(40, 130, n),
    "築年数": np.random.randint(0, 35, n),
    "駅徒歩": np.random.randint(1, 25, n)
})
data["価格"] = (
    35 * data["面積"] - 30 * data["築年数"]
    - 15 * data["駅徒歩"] + 500
    + np.random.randn(n) * 150
)

X = data[["面積", "築年数", "駅徒歩"]]
y = data["価格"]

# データの分割と学習
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42
)

model = LinearRegression()
model.fit(X_train, y_train)
y_pred = model.predict(X_test)

# 評価指標の計算
mse = mean_squared_error(y_test, y_pred)
rmse = np.sqrt(mse)
mae = mean_absolute_error(y_test, y_pred)
r2 = r2_score(y_test, y_pred)

print("【回帰モデルの評価結果】")
print(f"MSE  : {mse:>12.2f} (万円^2)")
print(f"RMSE : {rmse:>12.2f} (万円)")
print(f"MAE  : {mae:>12.2f} (万円)")
print(f"R^2  : {r2:>12.4f}")

import numpy as np
import pandas as pd
from sklearn.linear_model import LinearRegression, Ridge, Lasso
from sklearn.metrics import mean_squared_error
from sklearn.metrics import mean_absolute_error, r2_score

# 複数のモデルを比較
models = {
    "LinearRegression": LinearRegression(),
    "Ridge (alpha=1.0)": Ridge(alpha=1.0),
    "Ridge (alpha=10.0)": Ridge(alpha=10.0),
    "Lasso (alpha=1.0)": Lasso(alpha=1.0),
    "Lasso (alpha=10.0)": Lasso(alpha=10.0),
}

results = []
for name, m in models.items():
    m.fit(X_train, y_train)
    y_p = m.predict(X_test)
    results.append({
        "モデル": name,
        "RMSE": np.sqrt(mean_squared_error(y_test, y_p)),
        "MAE": mean_absolute_error(y_test, y_p),
        "R^2": r2_score(y_test, y_p)
    })

result_df = pd.DataFrame(results)
print(result_df.to_string(index=False))