PCA 手动实现（python）

import numpy as np
from sklearn.decomposition import PCA
def PCA_Demo(X, k):
    avg = np.mean(X, axis=1)
    for i in range(0, X.shape[1]):
        X[:, i] = X[:, i] - avg[i]
    print('X={}'.format(X))
    C = np.dot(np.transpose(X), X)
    a, b = np.linalg.eig(C)
    index = [[i, a[i]] for i in range(0, len(a))]
    sorted(index, key=lambda x: x[1])
    index_s = np.array([index[i][0] for i in range(0, k)])
    P = np.array([b[i] for i in index_s])
    return np.dot(X, P.T)

X = [[1, 2, 1], [1, 1, 2], [2, 1, 1], [5, 5, 5], [5, 5, 6], [5, 7, 5]]
y = [1, 1, 1, 0, 0, 0]

X_Demo = PCA_Demo(np.array(X), 2)



import numpy as np
X = np.array(X)
pca = PCA(n_components=2)
pca.fit(X)
X_pca = pca.transform(X)



from sklearn.svm import SVC
from sklearn.model_selection import train_test_split

x_train, x_test, y_train, y_test = train_test_split(X_Demo, y)
X_train, X_test, Y_train, Y_test = train_test_split(X_pca, y)

clf1 = SVC(kernel='rbf', class_weight='balanced')
clf1.fit(x_train, y_train)
pre_x = clf1.predict(x_test)
print(y_test)
print(pre_x)

clf1 = SVC(kernel='rbf', class_weight='balanced')
clf1.fit(X_train, Y_train)
pre_x = clf1.predict(X_test)
print(Y_test)
print(pre_x)