第十四次作业：手写数字识别-小数据集

1.手写数字数据集

# 1.手写数字数据集
from sklearn.datasets import load_digits
import numpy as np
digits = load_digits()  # 读取手写数字数据集

2.图片数据预处理

• x：归一化MinMaxScaler()

• # 对X_data进行归一化MinMaxScaler
  scaler = MinMaxScaler()
  X_data = scaler.fit_transform(X_data)
  print("X_data归一化后：",X_data

• 
• y：独热编码OneHotEncoder()或to_categorical
• 张量结构

• # OneHotEncoder独热编码
  from sklearn.preprocessing import OneHotEncoder

  y = digits.target.astype(np.float32).reshape(-1,1) #将Y_data变为一列
  Y = OneHotEncoder().fit_transform(y).todense() #张量结构todense
  print("进行Y独热编码：",Y)

• 
• 训练集测试集划分

• # 划分训练集和测试集
  from sklearn.model_selection import train_test_split
  x_train,x_test,y_train,y_test = train_test_split(X,Y,test_size=0.2,random_state=0,stratify=Y)
  print(x_train.shape, x_test.shape, y_train.shape, y_test.shape)

3.设计卷积神经网络结构

• 绘制模型结构图，并说明设计依据。

• #3.设计卷积神经网络结构
  from tensorflow.keras.models import Sequential
  from tensorflow.keras.layers import Dense,Dropout,Conv2D,MaxPool2D,Flatten
  #3、建立模型
  model = Sequential()
  ks = (3, 3)  # 卷积核的大小
  input_shape = x_train.shape[1:]
  # 一层卷积，padding='same',tensorflow会对输入自动补0
  model.add(Conv2D(filters=16, kernel_size=ks, padding='same', input_shape=input_shape, activation='relu'))
  # 池化层1
  model.add(MaxPool2D(pool_size=(2, 2)))
  # 防止过拟合，随机丢掉连接
  model.add(Dropout(0.25))
  # 二层卷积
  model.add(Conv2D(filters=32, kernel_size=ks, padding='same', activation='relu'))
  # 池化层2
  model.add(MaxPool2D(pool_size=(2, 2)))
  model.add(Dropout(0.25))
  # 三层卷积
  model.add(Conv2D(filters=64, kernel_size=ks, padding='same', activation='relu'))
  # 四层卷积
  model.add(Conv2D(filters=128, kernel_size=ks, padding='same', activation='relu'))
  # 池化层3
  model.add(MaxPool2D(pool_size=(2, 2)))
  model.add(Dropout(0.25))
  # 平坦层
  model.add(Flatten())
  # 全连接层
  model.add(Dense(128, activation='relu'))
  model.add(Dropout(0.25))
  # 激活函数softmax
  model.add(Dense(10, activation='softmax'))
  print(model.summary())

• 

    

4.模型训练

import matplotlib.pyplot as plt
# 画图
def show_train_history(train_history, train, validation):
    plt.plot(train_history.history[train])
    plt.plot(train_history.history[validation])
    plt.title('Train History')
    plt.ylabel('train')
    plt.xlabel('epoch')
    plt.legend(['train', 'validation'], loc='upper left')
    plt.show()
 
#4.模型训练
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
train_history = model.fit(x=x_train, y=y_train, validation_split=0.2, batch_size=300, epochs=10, verbose=2) #训练十次
# 准确率
show_train_history(train_history, 'accuracy', 'val_accuracy')
# 损失率
show_train_history(train_history, 'loss', 'val_loss')

 准确率：

 

 损失率：

          

5.模型评价

• model.evaluate()

• # 4、模型评价
  import pandas as pd
  import seaborn as sns
  score = model.evaluate(x_test,y_test)
  print("score:",score)
  # 预测值
  pre = model.predict_classes(x_test)
  print("预测值为：",pre[:10])

  

• 交叉表与交叉矩阵
• pandas.crosstab
• seaborn.heatmap

  # 交差表与交叉矩阵
  y_test1 = np.argmax(y_test,axis=1).reshape(-1)
  y_true = np.array(y_test1)[0]
  # 交叉表查看预测数据与原数据对比
  pd.crosstab(y_true,pre,rownames=['true'],colnames=['predict'])
  # 交叉矩阵
  y_test1 = y_test1.tolist()[0]
  a = pd.crosstab(np.array(y_test1),pre,rownames=['Lables'],colnames=['Predict'])
  # 转换成dataframe
  df = pd.DataFrame(a)
  sns.heatmap(df,annot=True,cmap="Oranges",linewidths=0.2,linecolor="G")