使用SKLearn中的SGD做数字识别

第一次调用已经写好的随机梯度下降算法！非常方便，之前还一直以为要自己写。数据都是SkLearn提供的，1700多张8*8的图片。最后只做了一个二分类。

import numpy as np
import os

import matplotlib
import matplotlib.pyplot as plt
plt.gray()
# plt.rcParams['axes.labelsize'] = 14
# plt.rcParams['xtick.labelsize'] = 12
# plt.rcParams['ytick.labelsize'] = 12

import warnings
warnings.filterwarnings('ignore')
np.random.seed(42)

from sklearn.datasets import load_digits
X, y = load_digits(return_X_y=True)

data_len = y.shape[0]
train_data_len = 1500
test_data_len = data_len - train_data_len

# 定义训练数据和测试数据
X_train, X_test = X[: train_data_len], X[train_data_len:]
y_train, y_test = y[: train_data_len], y[train_data_len:]

shuffle_index = np.random.permutation(train_data_len)
X_train, y_train = X_train[shuffle_index], y_train[shuffle_index]


y_train_6 = (y_train == 6)
y_test_6 = (y_test == 6)

from sklearn.linear_model import SGDClassifier
sgd_clf = SGDClassifier(max_iter=20, random_state=42)
sgd_clf.fit(X_train, y_train_6)

suspects = []
for i in range(train_data_len, data_len):
    if y[i] == 6:
        cur = sgd_clf.predict([X[i]])[0]
        if cur == False:
            suspects.append(i)

X = load_digits()

for i in suspects:
    plt.matshow(X.images[i])
plt.show()
print(len(suspects))