mnist-TSNE特征

import keras
import matplotlib.pyplot as plt
from sklearn.manifold import TSNE

(x_train,y_train), _ = keras.datasets.mnist.load_data()
x_train = x_train.reshape((x_train.shape[0],-1))

tsne = TSNE(n_components=2,random_state=0)
reduce_feat = tsne.fit_transform(x_train[:200])

fig = plt.figure('feature')
plt.scatter(reduce_feat[:,0], reduce_feat[:,1], c=y_train[:200],label="t-SNE")
for i in range(200):
    plt.text(reduce_feat[i,0],reduce_feat[i,1],str(y_train[i]),color='b')
plt.show()