全部八个步骤使用代码实现

以下是全部八个步骤的Python代码实现：

**步骤1：数据预处理**

import pandas as pd
from sklearn.preprocessing import StandardScaler

# 读取数据
df = pd.read_csv('white_wine.csv')

# 数据标准化
scaler = StandardScaler()
df_scaled = pd.DataFrame(scaler.fit_transform(df), columns=df.columns)
df_scaled.to_csv('white2_wine.csv', index=False)

**步骤2：划分训练集和测试集**

from sklearn.model_selection import train_test_split

# 划分数据
X = df_scaled.drop('quality', axis=1) # 特征
y = df_scaled['quality'] # 目标变量
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.1, random_state=42)

**步骤3：构建分类器**

from sklearn.linear_model import LogisticRegression
from sklearn.naive_bayes import GaussianNB
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.cluster import KMeans

classifiers = [
    ('LR', LogisticRegression()),
    ('NB', GaussianNB()),
    ('DT', DecisionTreeClassifier(random_state=42)),
    ('RF', RandomForestClassifier(random_state=42)),
    ('KM', KMeans(n_clusters=2))
]

for name, clf in classifiers:
    model = clf.fit(X_train, y_train)
    # ... (继续保存模型或进行后续操作)

**步骤4：PCA降维**

from sklearn.decomposition import PCA

pca = PCA(n_components=3)
X_train_pca = pca.fit_transform(X_train)
X_test_pca = pca.transform(X_test)

# 保存降维后的数据
df_pca = pd.DataFrame(X_train_pca, columns=['PC1', 'PC2', 'PC3'])
df_pca.to_csv('whine3_wine.csv', index=False)

**步骤5：在降维后建模并评估**

for name, clf in classifiers:
    model = clf.fit(X_train_pca, y_train)
    accuracy = clf.score(X_test_pca, y_test)
    print(f"{name}在降维后测试集准确率为：{accuracy}")

# 可视化准确率
plt.figure(figsize=(10, 6))
plt.bar(classifiers.keys(), [accuracy for _, _ in classifiers], label='Accuracy')
plt.xlabel('Classifier')
plt.ylabel('Accuracy')
plt.title('PCA后各分类器测试集准确率')
plt.legend()

**步骤6：随机验证数据**

random_data = df_scaled.sample(10).reset_index(drop=True)
print(random_data)

**步骤7：模型预测及可视化**

for name, clf in classifiers:
    predictions = clf.predict(random_data.drop('quality', axis=1))
    visualize_predictions(predictions, random_data['quality'], name)

**步骤8：降维后预测**

for name, clf in classifiers:
    predictions_pca = clf.predict(X_test_pca)
    visualize_predictions_pca(predictions_pca, y_test, name)

这里假设`visualize_predictions`和`visualize_predictions_pca`是自定义的函数，用于可视化预测结果。记得替换它们的实际实现。