지도학습(K-NN) 과 비지도학습(K-Means) 비교

1. 지도학습 : K-NN

# iris dataset 으로 분류모델 작성 : 지도학습(K-NN), 비지도학습(K-Means)

from sklearn.datasets import load_iris

iris = load_iris()

from sklearn.model_selection import train_test_split
train_x, test_x, train_y, test_y = train_test_split(iris['data'], iris['target'],
                                                    test_size=0.25, random_state=42)

print(train_x[:2])
print(train_y[:2])

from sklearn.neighbors import KNeighborsClassifier
print('지도학습 : K최근접이웃알고리즘')
knnModel = KNeighborsClassifier(n_neighbors = 3)
knnModel.fit(train_x, train_y) # feature, label

predict_label = knnModel.predict(test_x)
print(predict_label[:3])

from sklearn import metrics
print('acc:',metrics.accuracy_score(test_y, predict_label))

2. 비지도학습 : k-means

print('비지도학습 : K평균 군집 알고리즘')
from sklearn.cluster import KMeans
kmeansModel = KMeans(n_clusters = 3, init='k-means++', random_state=0)
kmeansModel.fit(train_x) # label 이 없다

print(kmeansModel.labels_)
print('0 cluster:', train_y[kmeansModel.labels_ == 0])
print('1 cluster:', train_y[kmeansModel.labels_ == 1])
print('2 cluster:', train_y[kmeansModel.labels_ == 2])

print()
import numpy as np
new_input = np.array([[1.1, 2.3, 1.5, 1.5]])
clu_pred = kmeansModel.predict(new_input)
print(clu_pred)