import tensorflow as tf
import numpy as np
import pandas as pd
#数据预处理
def read_data():
data=pd.read_csv('train.csv') #pandas 读取
data=data.fillna(0) #NAN 填入0
datax=data[['Sex', 'Age', 'Pclass', 'SibSp', 'Parch', 'Fare','Embarked']] #pandas选择列
datax=pd.get_dummies(datax) # one—hot 编码
data['Deceased'] = data['Survived'].apply(lambda s: 1 - s) #one -hot编码
datay = data[['Survived','Deceased']]
return datax,datay
if __name__=="__main__":
new_datax,new_datay=read_data()
new_dataxx = new_datax.values.astype(np.float32) #把pandas矩阵 转化为np矩阵没有二维标签变为存矩阵
new_datayy = new_datay.values.astype(np.float32) # 把pandas矩阵 转化为np矩阵没有二维标签变为存矩阵
np.random.shuffle(new_dataxx) #随机打乱顺序
np.random.shuffle(new_datayy) # 随机打乱顺序
sep = int(0.7 * len(new_dataxx)) #拆分数据集
train_datax = new_dataxx[:sep] # training data (70%)
train_datay=new_datayy[:sep]
test_datax = new_dataxx[sep:] # test data (30%)
test_datay=new_datayy[sep:]
# new_datax.to_csv("Taitan_onehot.csv", index=False)
#简便方法 打乱+拆分
'''
分割的简单方法 from sklearn.model_selection import train_test_split
X_train, X_val, y_train, y_val = train_test_split(dataset_X, dataset_Y,
test_size=0.2,
random_state=42)
'''
#全链接神经层 的输入 tf_inputx 训练数据的特征信息 tf_inputy训练数据的标签
tf_inputx=tf.placeholder(tf.float32, [None, 11])
tf_inputy=tf.placeholder(tf.float32,[None,2])
#搭建全链接神经网络
l1 = tf.layers.dense(tf_inputx, 1000, tf.nn.relu, name="l1")
l2 = tf.layers.dense(l1, 900, tf.nn.relu, name="l2")
out = tf.layers.dense(l2, 2, name="l3")
#计算概率 相加为1
prediction=tf.nn.softmax(out, name="pred")
#计算误差 代价函数
loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=tf_inputy, logits=out))
#优化器 减少误差
train_step = tf.train.AdamOptimizer(0.001).minimize(loss)
#初始化变量
init = tf.global_variables_initializer()
'''
correct_prediction = tf.equal(tf.argmax(tf_inputy, 1), tf.argmax(prediction, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
'''
#创建会话
with tf.Session() as sess:
sess.run(init) #初始化变量
for i in range(2000): #控制迭代次数
# 启动优化其 减少误差 开始训练
sess.run(train_step, feed_dict={tf_inputx: train_datax, tf_inputy: train_datay})
# 查看当前训练实时数据反馈
print("number:"+str(i))
print(sess.run(out, feed_dict={tf_inputx: train_datax, tf_inputy: train_datay}))
# acc = sess.run(accuracy, feed_dict={tf_inputx: test_datax, tf_inputy: test_datay})
# print("Iter" + str(i) + ",Testing Accuracy" + str(acc))
#训练完成后 进行预测 查看预测结果
print("test")
print(sess.run(prediction,feed_dict={tf_inputx:test_datax}))
