深度学习基于Keras的Python实践——印第安糖尿病人诊断

数据集下载
csdn资源:https://download.csdn.net/download/qq_35008279/11122908
百度云链接:https://pan.baidu.com/s/1duoc2qaEyowuwWUs7gRs8Q 密码:pe9y

from keras.models import Sequential
from keras.layers import Dense
import numpy as np
# 设定随机数种子
np.random.seed(7)

# 导入数据
dataset = np.loadtxt('amp_full_num21.csv', delimiter=',')
# 分割输入x和输出Y
x = dataset[:, 1 : 200]
Y = dataset[:, 0]

# 创建模型
model = Sequential()
model.add(Dense(12, input_dim=199, activation='relu'))
model.add(Dense(8, activation='relu'))
model.add(Dense(1, activation='sigmoid'))

# 编译模型
model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])

# 训练模型
model.fit(x=x, y=Y, epochs=150, batch_size=10)

# 评估模型
scores = model.evaluate(x=x, y=Y)
print('\n%s : %.2f%%' % (model.metrics_names[1], scores[1]*100))

结果