15 手写数字识别-小数据集

1 from sklearn.datasets import load_digits
2 digits = load_digits()

 1 #对x集归一化;对y进行热编码
 2 import matplotlib.pyplot as plt
 3 import numpy as np
 4 from sklearn.datasets import load_digits  #小数据集为8*8/大的为28*28
 5 from sklearn.model_selection import train_test_split
 6 from sklearn.preprocessing import MinMaxScaler
 7 from sklearn.preprocessing import OneHotEncoder
 8 import tensorflow as tf
 9 from sklearn.metrics import accuracy_score
10 
11 X_data = digits.data.astype(np.float32)
12 Y_data = digits.target.astype(np.float32).reshape(-1,1) #将数据转为一列
13 
14 #将属性xSHUJHU数据进行归一放在最大与最小之间（0,1）
15 scaler = MinMaxScaler()
16 X_data = scaler.fit_transform(X_data)  #归一
17 print("看效果",X_data)

from sklearn.preprocessing import OneHotEncoder
Y = OneHotEncoder().fit_transform(Y_data).todense()
print(‘one-hot_Y:‘)
print(Y)   # 热编码有一说一

1 #先对归一的数据集转为图片格式
2 X=X_data.reshape(-1,8,8,1)
3 
4 #训练集测试集划分
5 from sklearn.model_selection import train_test_split  #测试与训练2/8分
6 X_tain,X_test,y_train,y_test = train_test_split(X,Y,test_size=0.2,random_state=0,stratify=Y)
7 print(X_tain.shape,X_test.shape,y_train.shape,y_test.shape)

from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense,Dropout,Conv2D,MaxPool2D,Flatten
model = Sequential() # 建立模型
#设定卷积核
ks = (2, 2)
#输入层（1层）此处input_shape需要指定训练集的数据，往下之后是会自动推导
model.add(Conv2D(filters=8, kernel_size=ks, padding=‘same‘, input_shape=X_tain.shape[1:], activation=‘relu‘))
model.add(MaxPool2D(pool_size=(2, 2))) # 池化层
model.add(Dropout(0.25)) # 防止过拟合丢掉1/4的链接

model.add(Conv2D(filters=16, kernel_size=ks, padding=‘same‘, activation=‘relu‘))# 第2卷积层
model.add(MaxPool2D(pool_size=(2, 2)))# 池化
model.add(Dropout(0.25))# 防止过拟合丢掉1/4的链接

model.add(Conv2D(filters=324, kernel_size=ks, padding=‘same‘, activation=‘relu‘))# 第三卷积层
model.add(Conv2D(filters=64, kernel_size=ks, padding=‘same‘, activation=‘relu‘))# 第四卷积层

model.add(MaxPool2D(pool_size=(2, 2)))# 池化
model.add(Dropout(0.25))

model.add(Flatten())# 平坦层

model.add(Dense(64, activation=‘relu‘))# 全连接层
model.add(Dropout(0.25))

model.add(Dense(10, activation=‘softmax‘))# 激活函数softmax
# 输出模型各层的参数状况
print(model.summary())