머신 러닝 (10) 비지도 PCA

 import numpy as np

import pandas as pd

import matplotlib.pyplot as plt

import koreanize_matplotlib

import seaborn as sns

from sklearn.datasets import load_diabetes

from sklearn.model_selection import train_test_split, cross_val_score

from sklearn.preprocessing import StandardScaler

from sklearn.decomposition import PCA

import warnings

warnings.filterwarnings('ignore')

# Feature(X)와 Target(y) 분리

diabetes = load_diabetes()

X = diabetes.data

y = diabetes.target

scaler = StandardScaler()

X_scaled = scaler.fit_transform(X)

print(f"스케일링된 데이터 형태: {X_scaled.shape}")

# 2개의 주성분으로 축소

pca = PCA(n_components=X_scaled.shape[1])

X_pca = pca.fit_transform(X_scaled)

print(f"PCA 변환 후 차원: {X_pca.shape}")

print(f"주성분별 설명된 분산 비율: {pca.explained_variance_ratio_}")

print(f"총 설명된 분산 비율: {np.sum(pca.explained_variance_ratio_):.4f}")

# 전체 주성분에 대한 PCA 수행

pca_full = PCA(n_components=X_scaled.shape[1])

pca_full.fit(X_scaled)

# 설명된 분산 비율 및 누적 분산 비율 계산

explained_variance = pca_full.explained_variance_ratio_

cumulative_variance = np.cumsum(explained_variance)

# 2개의 주성분으로 축소

pca = PCA(n_components=5)

X_pca = pca.fit_transform(X_scaled)

print(f"PCA 변환 후 차원: {X_pca.shape}")

print(f"주성분별 설명된 분산 비율: {pca.explained_variance_ratio_}")

print(f"총 설명된 분산 비율: {np.sum(pca.explained_variance_ratio_):.4f}")

# Scree Plot 시각화

plt.figure(figsize=(8, 5))

plt.bar(range(1, len(explained_variance) + 1), explained_variance, alpha=0.5, align='center',

        label='Individual explained variance')

plt.step(range(1, len(cumulative_variance) + 1), cumulative_variance, where='mid',

         label='Cumulative explained variance')

plt.ylabel('Explained variance ratio')

plt.xlabel('Principal component index')

plt.title('Scree Plot')

plt.legend(loc='best')

plt.grid(True)

plt.show()

plt.figure(figsize=(6, 4))

# Outcome(당뇨병 발병 여부)에 따라 색상 구분 (0: 정상, 1: 당뇨)

scatter = plt.scatter(X_pca[:, 0], X_pca[:, 1], c=y, cmap='viridis', alpha=0.7, edgecolors='k')

plt.xlabel('Principal Component 1')

plt.ylabel('Principal Component 2')

plt.title('PCA of Diabetes Dataset')

plt.colorbar(scatter, label='Diabetes Target Value')

plt.grid(True)

plt.show()