基于CNN的鸢尾花分类器

基于CNN的鸢尾花分类器

鸢尾花为结构化数据，包括四个特征，三个标签

本次使用CNN来进行分类

 

废话不多说，请看代码

 

#导入所要用到的包

from sklearn import datasets
from keras.layers import *
import keras.backend as K
from keras import Model
from keras.utils import plot_model,to_categorical
import pandas as pd
import numpy as np
from keras.utils import plot_model
from sklearn.utils import shuffle
import matplotlib.pyplot as plt

 

#划分特征和标签

iris = datasets.load_iris()
#iris = np.loadtxt('./iris.csv', delimiter='\t')
X = np.array(iris['data'])
Y = np.array(iris['target'])

fea_cnt = len(X[0]) #特征个数
numb = max(Y)+1 # 分类类别个数

Y = np.array([to_categorical(y,numb,dtype='int32') for y in Y])

 

#构建模型，二维卷积
def build_model(fea_cnt,numb):
    K.clear_session() #清除之前的模型，省得压满内存
    inputs = Input(shape=(fea_cnt,), dtype='float32')
    embds = Dense(64 ,activation='selu')(inputs)
    embds = Reshape((8,8,1))(embds)
    embds = Conv2D(1,(4,4),strides=(1,1),padding="valid")(embds)
    embds = MaxPooling2D(pool_size=(2,2), strides=None, padding='valid')(embds)
    embds = Flatten()(embds)
    embds = Dense(64,activation='selu')(embds)
    embds = Dense(32,activation='selu')(embds)
    outputs = Dense(numb,activation='softmax')(embds)
    
    model = Model(inputs=inputs, outputs=outputs)
    model.compile(loss='categorical_crossentropy', optimizer = 'adam', metrics=['accuracy'])
    #model.summary()
    return model

#数据集划分，20%的数据为测试集

model = build_model(fea_cnt,numb)
X,Y = shuffle(X,Y, random_state=1337) #打乱顺序进行分隔
validation_split=0.2 

train_X = X[:int(len(X)*(1-validation_split))]
train_Y = Y[:int(len(Y)*(1-validation_split))] 
test_X = X[int(len(X)*(1-validation_split)):]
test_Y = Y[int(len(Y)*(1-validation_split)):]

#训练模型
epochs = 12 
batch_size = 4
model.fit(train_X ,train_Y
              ,batch_size = batch_size
              ,epochs = epochs
              ,validation_data= [test_X,test_Y]
        )

plot_model(model, to_file='./model.png',show_shapes=True,dpi=300) #输出框架图
score = model.evaluate(test_X, test_Y)

#打印准确度与损失值

print('Test loss:', score[0])
print('Test accuracy:', score[1])