神经网络load_wine()#调库

#调库三部曲
#1.导包
import numpy as np
from sklearn.neural_network import MLPClassifier
import sklearn.datasets as dts
#2.给对象
data=dts.load_wine()
#10.获取特征x
x=data.data
#11.获取标签y
y=data.target
np.random.seed(5)
order=np.random.permutation(x.shape[0])
x=x[order]
y=y[order]
num=int((x.shape[0])*0.8)
train_x,test_x=np.split(x,[num])
train_y,test_y=np.split(y,[num])
#声明模型,给对象
model=MLPClassifier(hidden_layer_sizes=(9,8),max_iter=100,alpha=0.1)
#3.训练
model.fit(train_x,train_y)
#测试集精度
print('测试集精度:',model.score(test_x,test_y))
#预测测试集
h=model.predict(test_x)
print('预测测试集:\n',h)
# 计算精度f1
f1=model.score(train_x, train_y)
print('精度f1:',f1)
#输出类别list
print('类别list:',model.classes_)
#当前的loss值
print('当前的loss值:',model.loss_)
# list,权重项
print('list权重项:',model.coefs_)
# list,偏置项
print('list偏置项:',model.intercepts_)
#solver已经运行的迭代次数
print('solver已经运行的迭代次数:',model.n_iter_)
# NN层数
print('NN层数:',model.n_layers_)
# 输出个数
print('输出个数:',model.n_outputs_)
print('输出激活函数的名字:',model.out_activation_)