In [1]:
%matplotlib inline
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
In [2]:
digits_traiin = pd.read_csv('https://archive.ics.uci.edu/ml/machine-learning-databases/optdigits/optdigits.tra', header=None)
In [3]:
digits_test = pd.read_csv('https://archive.ics.uci.edu/ml/machine-learning-databases/optdigits/optdigits.tes', header=None)
In [4]:
X_train = digits_traiin[np.arange(64)]
Y_train = digits_traiin[64]
In [5]:
X_tesy = digits_test[np.arange(64)]
Y_test = digits_test[64]
In [6]:
#显示手写数字图片的特征经过PCA压缩后的二维空间分布
from sklearn.decomposition import PCA
In [7]:
#初始化一个可以讲高纬度特征向量（六十四维）压缩到2个纬度的PCA
estimator = PCA(n_components=2)
In [8]:
X_pca = estimator.fit_transform(X_train)
In [9]:
def plot_pca_scatter():
colors = ['black','blue','purple','yellow','white','red','lime','cyan','orange','gray']
for i in range(len(colors)):
px = X_pca[:,0][Y_train.as_matrix()==i]
py = X_pca[:,1][Y_train.as_matrix()==i]
plt.scatter(px, py, c = colors[i])
plt.legend(np.arange(0,10).astype(str))
plt.xlabel('First Principal Component')
plt.ylabel('Second Principal Component')
plt.show()
plot_pca_scatter()
In [10]:
#使用压缩后的未读进行训练，取20个特征
estimator = PCA(n_components=20)
In [11]:
pca_X_train = estimator.fit_transform(X_train)
In [12]:
pca_X_test = estimator.transform(X_tesy)
In [13]:
from sklearn.svm import LinearSVC
In [14]:
svc = LinearSVC()
svc.fit(X_train, Y_train)
y_pred = svc.predict(X_tesy)
print y_pred
In [15]:
pca_svc = LinearSVC()
In [16]:
pca_svc.fit(pca_X_train, Y_train)
Out[16]:
In [17]:
pca_y_pred = pca_svc.predict(pca_X_test)
In [18]:
from sklearn.metrics import classification_report
In [19]:
y_pred = y_pred.reshape(-1,1)
pca_y_pred = pca_y_pred.reshape(-1,1)
print svc.score(X_tesy, Y_test)
print classification_report(Y_test, y_pred, target_names=np.arange(10).astype(str))
print pca_svc.score(pca_X_test, Y_test)
print classification_report(Y_test, pca_y_pred, target_names=np.arange(10).astype(str))