通过sklearn实现简易手写数字识别

本测试在jupyter中实现
直接上代码，入门级别手写数字识别：

import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
from sklearn.neighbors import KNeighborsClassifier

#使用了1500个数据，比较少，相关的data图片可以在文章后面进行下载
data = []
for i in range(10):
    for j in range(1,151):
    #使用'./data/%d/%d_%d.bmp'定位到文件
        data.append(plt.imread('./data/%d/%d_%d.bmp'%(i,i,j)))

y = [0,1,2,3,4,5,6,7,8,9]*150
y = np.array(y)
y = y.sort()

#1500个图片中，任选1000作为训练样本，500作为测试样本，可能会有重复，但是影响不大
index = np.random.randint(0,1500,size = 1000)
x_train = x[index]
y_train = y[index]
index = np.random.randint(0,1500,size = 500)
x_test = x[index]
y_test = y[index]

knn = KNeighborsClassifier(n_neighbors=5)
knn.fit(x_train.reshape(1000,784),y_train)

knn.predict(x_test.reshape(500,784))

y_ = knn.predict(x_test.reshape(500,784))

y_[:20]

y_test[:20]

(y_test == y_).mean()

data文件网盘链接：
https://pan.baidu.com/s/13f9vuxWN5WqyoGZkMf1qNQ