from sklearn.datasets import load_breast_cancer
import pandas as pd

# Load the dataset
cancer = load_breast_cancer()
X_cancer = pd.DataFrame(cancer.data, columns=cancer.feature_names)
y_cancer = pd.Series(cancer.target).map({0: '恶性', 1: '良性'})

print('数据集维度 (样本数, 特征数):', X_cancer.shape)
print('\n前5行数据预览:')
# Display the head of the DataFrame for a quick look at the features
with pd.option_context('display.max_columns', 5, 'display.width', 100):
    print(X_cancer.head())

from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler

# 1. Standardize the features to have mean=0 and variance=1
scaler = StandardScaler()
X_scaled = scaler.fit_transform(X_cancer)

# 2. Split data into training and testing sets (70% train, 30% test)
# stratify=y_cancer 确保训练集和测试集中的类别比例与原始数据集相同
X_train, X_test, y_train, y_test = train_test_split(
    X_scaled, y_cancer, test_size=0.3, random_state=42, stratify=y_cancer
)

print(f'训练集大小: {X_train.shape} 样本')
print(f'测试集大小: {X_test.shape} 样本')

from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import accuracy_score

# 初始化分类器，并设置 k=5
knn_clf = KNeighborsClassifier(n_neighbors=5)

# 在训练数据上训练模型
knn_clf.fit(X_train, y_train)

# 在未见过的测试数据上进行预测
y_pred = knn_clf.predict(X_test)

# 计算准确率: (正确预测数) / (总预测数)
accuracy = accuracy_score(y_test, y_pred)
print(f'当 k=5 时，模型在测试集上的准确率为: {accuracy:.4f}')

import matplotlib.pyplot as plt
import numpy as np
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import accuracy_score

# --- Recalculate from previous cell ---
from sklearn.datasets import load_breast_cancer
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
cancer = load_breast_cancer()
y_cancer = pd.Series(cancer.target)
scaler = StandardScaler()
X_scaled = scaler.fit_transform(cancer.data)
X_train, X_test, y_train, y_test = train_test_split(
    X_scaled, y_cancer, test_size=0.3, random_state=42, stratify=y_cancer
)
# --- End recalculate ---

# 我们将测试 k 值从 1 到 30
k_range = range(1, 31)
test_scores = []

# 遍历每个 k 值
for k_val in k_range:
    knn_loop = KNeighborsClassifier(n_neighbors=k_val)
    knn_loop.fit(X_train, y_train)
    y_pred_loop = knn_loop.predict(X_test)
    test_scores.append(accuracy_score(y_test, y_pred_loop))

# 找到准确率最高的 k 值
best_k = k_range[np.argmax(test_scores)]
best_accuracy = max(test_scores)

# --- Visualization ---

plt.figure(figsize=(10, 6))
plt.plot(k_range, test_scores, marker='o', linestyle='--', color='royalblue', mfc='skyblue')
plt.title(f'模型准确率随k值的变化', fontsize=18, pad=15)
plt.xlabel('k值 (邻居数量)', fontsize=14)
plt.ylabel('测试集准确率', fontsize=14)
plt.axvline(best_k, color='red', linestyle=':', linewidth=2, label=f'最优 k = {best_k}\n准确率 = {best_accuracy:.4f}')
plt.xticks(np.arange(0, 31, 2))
plt.grid(True, linestyle='--', alpha=0.6)
plt.legend(fontsize=12)
plt.tight_layout()
plt.show()

# --- Recalculate from previous cells to ensure context ---
from sklearn.datasets import load_breast_cancer
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import accuracy_score
import pandas as pd
cancer = load_breast_cancer()
y_cancer = pd.Series(cancer.target)
scaler = StandardScaler()
X_scaled = scaler.fit_transform(cancer.data)
X_train, X_test, y_train, y_test = train_test_split(
    X_scaled, y_cancer, test_size=0.3, random_state=42, stratify=y_cancer
)
# --- End recalculate ---

# 我们从上一步骤得知最优k为7
best_k = 7 

# 训练一个标准KNN模型 (权重统一)
knn_standard = KNeighborsClassifier(n_neighbors=best_k, weights='uniform')
knn_standard.fit(X_train, y_train)
accuracy_standard = accuracy_score(y_test, knn_standard.predict(X_test))

# 训练一个加权KNN模型 (权重基于距离)
knn_weighted = KNeighborsClassifier(n_neighbors=best_k, weights='distance')
knn_weighted.fit(X_train, y_train)
accuracy_weighted = accuracy_score(y_test, knn_weighted.predict(X_test))

print(f'标准KNN (k={best_k}) 准确率: {accuracy_standard:.4f}')
print(f'加权KNN (k={best_k}) 准确率: {accuracy_weighted:.4f}')

import pandas as pd
import numpy as np
from sklearn.datasets import fetch_california_housing
from sklearn.neighbors import KNeighborsRegressor
from sklearn.metrics import mean_squared_error
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler

# 加载并准备数据
housing = fetch_california_housing()
X_h = pd.DataFrame(housing.data, columns=housing.feature_names)
y_h = pd.Series(housing.target) # 目标单位是 $100,000

# 特征标准化和数据划分
scaler_h = StandardScaler()
X_h_scaled = scaler_h.fit_transform(X_h)
X_h_train, X_h_test, y_h_train, y_h_test = train_test_split(X_h_scaled, y_h, test_size=0.3, random_state=42)

# 训练KNN回归模型，k=7
knn_reg = KNeighborsRegressor(n_neighbors=7)
knn_reg.fit(X_h_train, y_h_train)
y_h_pred = knn_reg.predict(X_h_test)

# 使用均方根误差(RMSE)进行评估
rmse = np.sqrt(mean_squared_error(y_h_test, y_h_pred))
print(f'KNN回归模型在测试集上的RMSE为: ${rmse * 100000:,.2f}')