Machine learning 13 (pca)

PCA

붓꽃 데이터

import pandas as pd
import matplotlib.pyplot as plt
from sklearn.datasets import load_iris

iris = load_iris()

iris_df = pd.DataFrame(iris.data, columns=iris.feature_names)
iris_df['target'] = iris.target
iris_df.head(3)

	sepal length (cm)	sepal width (cm)	petal length (cm)	petal width (cm)	target
0	5.1	3.5	1.4	0.2	0
1	4.9	3.0	1.4	0.2	0
2	4.7	3.2	1.3	0.2	0

sepal_length, sepal_width 두개의 특성으로 산점도

plt.scatter(iris_df['sepal length (cm)'], iris_df['sepal width (cm)'], c=iris_df['target'])
plt.xlabel('sepal length')
plt.ylabel('sepal width')
plt.show()

X = iris.data
y = iris.target

from sklearn.preprocessing import StandardScaler

scaler = StandardScaler()
iris_scaled = scaler.fit_transform(X)

from sklearn.decomposition import PCA

pca = PCA(n_components=2) # 4차원을 특성을 2차원으로 축소
iris_pca = pca.fit_transform(iris_scaled)

iris_pca.shape

(150, 2)

pca_columns = ['pca_component1', 'pca_component2']
iris_pca_df = pd.DataFrame(iris_pca, columns=pca_columns)
iris_pca_df['target'] = iris.target
iris_pca_df.head(3)

	pca_component1	pca_component2	target
0	-2.264703	0.480027	0
1	-2.080961	-0.674134	0
2	-2.364229	-0.341908	0

PCA로 차원축소된 새로운 특성으로 산점도

plt.scatter(iris_pca_df['pca_component1'], iris_pca_df['pca_component2'], c=iris_pca_df['target'])
plt.xlabel('pca component 1')
plt.ylabel('pca component 2')
plt.show()

pca.explained_variance_ratio_ # 각 pca component 의 분산 비율

array([0.72962445, 0.22850762])

0.72962445 + 0.22850762

0.95813207

원본 데이터와 PCA 변환된 데이터로 모델 성능 측정

from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import cross_val_score

rf_clf = RandomForestClassifier(random_state=42)
scores = cross_val_score(rf_clf, X, y, scoring='accuracy', cv=3)
scores

array([0.98, 0.94, 0.98])

rf_clf = RandomForestClassifier(random_state=42)
scores = cross_val_score(rf_clf, iris_pca, y, scoring='accuracy', cv=3)
scores

array([0.88, 0.88, 0.9 ])

iris_pca.shape

(150, 2)

Reference

• 파이썬 머신러닝 완벽 가이드 (권철민 저)