主成分分析 (PCA) 可视化和特征重要性分析

import pandas as pd
from sklearn.decomposition import PCA
import matplotlib.pyplot as plt
import numpy as np

# 读取Excel表格
data = pd.read_excel('C:\Users\lenovo\Desktop\HIV\GSE6740GSE50011基因降低\data1.xlsx')
X = data.iloc[:, 1:].values # 影响因素
y = data.iloc[:, 0].values # 因变量

# 主成分分析算法
pca = PCA(n_components=10)
pca.fit(X)
X_pca = pca.transform(X)


# 散点图
plt.scatter(X_pca[:, 0], X_pca[:, 1])
plt.title('PCA Scatter Plot')
plt.xlabel('PC1')
plt.ylabel('PC2')
plt.show()

# 热力图
corr = np.corrcoef(X_pca.T)
plt.imshow(corr, cmap='hot', interpolation='nearest')
plt.title('PCA Heatmap')
plt.colorbar()
plt.show()

# 3D图
from mpl_toolkits.mplot3d import Axes3D
fig = plt.figure()
ax = Axes3D(fig)
ax.scatter(X_pca[:, 0], X_pca[:, 1], X_pca[:, 2])
ax.set_xlabel('PC1')
ax.set_ylabel('PC2')
ax.set_zlabel('PC3')
plt.title('PCA 3D Plot')
plt.show()

# 折线图
plt.plot(X_pca)
plt.title('PCA Line Plot')
plt.xlabel('Samples')
plt.ylabel('Feature Values')
plt.show()

# 柱状图
plt.bar(range(len(pca.explained_variance_ratio_)), pca.explained_variance_ratio_)
plt.title('PCA Explained Variance Ratio')
plt.xlabel('Principal Component')
plt.ylabel('Variance Ratio')
plt.show()

# 输出方差贡献率和累计贡献率
print('Explained Variance Ratio:')
print(pca.explained_variance_ratio_)
print('Cumulative Explained Variance Ratio:')
print(np.cumsum(pca.explained_variance_ratio_))

# 绘制累计贡献率的折线图
plt.plot(np.cumsum(pca.explained_variance_ratio_))
plt.title('PCA Cumulative Explained Variance Ratio')
plt.xlabel('Principal Component')
plt.ylabel('Cumulative Variance Ratio')
plt.show()

# 输出影响因素的重要性
important_features = pd.DataFrame(pca.components_, columns=data.columns[1:])
important_features = important_features.abs().sum(axis=0).sort_values(ascending=False)
print('Feature Importance:')
print(important_features)


# 输出每个主成分的元素
for i in range(10):
    components = pd.DataFrame(pca.components_[i], index=data.columns[1:], columns=['Component ' + str(i+1)])
    components = components.abs().sort_values(by=['Component ' + str(i+1)], ascending=False)
    print('Top 10 elements in Component ' + str(i+1) + ':')
    print(components.head(10))