import torch
from torch import nn
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from sklearn.preprocessing import MinMaxScaler
from torch.nn import MaxPool2d, Conv2d, Dropout, ReLU
from torch.utils.data import DataLoader, Dataset
#准备数据集
df=pd.read_csv("train.csv",parse_dates=["Date"],index_col=[0])
print(df.shape)
train_data_size=round(len(df)*0.8)
test_data_size=round(len(df)*0.2)
print("训练数据集的长度为：{}".format(train_data_size))
print("测试数据集的长度为：{}".format(test_data_size))
df[['Open']].plot()
plt.ylabel("stock price")
plt.xlabel("times")
plt.show()
sel_col = ['Open', 'High', 'Low', 'Close']
df=df[sel_col]
df_close_max=df['Close'].max()
df_close_min=df['Close'].min()
print("最高价=", df_close_max)
print("最低价=", df_close_min)
print("波动值=", df_close_max-df_close_min)
print("上涨率=", (df_close_max-df_close_min)/df_close_min)
print("下跌率=", (df_close_max-df_close_min)/df_close_max)
df=df.apply(lambda x:(x-min(x))/(max(x)-min(x)))
print(df)
total_len=df.shape[0]
print("df.shape=",df.shape)
print("df_len=", total_len)
sequence=10
x=[]
y=[]
for i in range(total_len-sequence):
x.append(np.array(df.iloc[i:(i+sequence),].values,dtype=np.float32))
y.append(np.array(df.iloc[(i+sequence),1],dtype=np.float32))
print("train data of item 0: \n", x[0])
print("train label of item 0: \n", y[0])
print("\n序列化后的数据形状：")
X = np.array(x)
Y = np.array(y)
Y = np.expand_dims(Y, 1)
print("X.shape =",X.shape)
print("Y.shape =",Y.shape)
x_tensor=torch.from_numpy(X)
y_tensor=torch.from_numpy(Y)
train_x = x_tensor[:int(0.7 * total_len)]
train_y = y_tensor[:int(0.7 * total_len)]
# 数据集前70%后的数据（30%）作为验证集
valid_x = x_tensor[int(0.7 * total_len):]
valid_y = y_tensor[int(0.7 * total_len):]
print("训练集x的形状是：",train_x.shape)
print("测试集y的形状是：",train_y.shape)
print("测试集x的形状是：",valid_x.shape)
print("测试集y的形状是：",valid_y.shape)
class Mydataset(Dataset):
def __init__(self, x, y, transform=None):
self.x = x
self.y = y
def __getitem__(self, index):
x1 = self.x[index]
y1 = self.y[index]
return x1, y1
def __len__(self):
return len(self.x)
dataset_train = Mydataset(train_x, train_y)
dataset_valid = Mydataset(valid_x, valid_y)
train_dataloader=DataLoader(dataset_train,batch_size=64)
valid_dataloader=DataLoader(dataset_valid,batch_size=64)
class cnn_lstm(nn.Module):
def __init__(self,window_size,feature_number):
super(cnn_lstm, self).__init__()
self.window_size=window_size
self.feature_number=feature_number
self.conv1 = Conv2d(in_channels=1, out_channels=64, kernel_size=3, stride=1, padding=1)
self.relu1 = ReLU()
self.maxpooling1 = MaxPool2d(3, stride=1,padding=1)
self.dropout1 = Dropout(0.3)
self.lstm1 = nn.LSTM(input_size=64 * feature_number, hidden_size=128, num_layers=1, batch_first=True)
self.lstm2 = nn.LSTM(input_size=128, hidden_size=64, num_layers=1, batch_first=True)
self.fc = nn.Linear(in_features=64, out_features=32)
self.relu2 = nn.ReLU()
self.head = nn.Linear(in_features=32, out_features=1)
def forward(self, x):
x = x.reshape([x.shape[0], 1, self.window_size, self.feature_number])
# x = x.transpose(-1, -2)
x = self.conv1(x)
x = self.relu1(x)
x = self.maxpooling1(x)
x = self.dropout1(x)
x = x.reshape([x.shape[0], self.window_size, -1])
# x = x.transpose(-1, -2) #
x, (h, c) = self.lstm1(x)
x, (h, c) = self.lstm2(x)
x = x[:, -1, :] # 最后一个LSTM只要窗口中最后一个特征的输出
x = self.fc(x)
x = self.relu2(x)
x = self.head(x)
return x
#创建网络模型
cnn_lstm=cnn_lstm(window_size=10,feature_number=4)
print(cnn_lstm)
#定义损失函数
loss_fn=nn.MSELoss(size_average=True)
#定义优化器
learning_rate=0.01
opitmizer=torch.optim.Adam(cnn_lstm.parameters(),learning_rate)
#设置训练网络参数
total_train_step=0
total_valid_step=0
#训练论数
epoch=100
hist = np.zeros(epoch)
for i in range(epoch):
#print("______第{}轮训练开始________".format((i + 1)))
y_train_pred=cnn_lstm(train_x)
loss=loss_fn(y_train_pred,train_y)
if i % 10 == 0 and i != 0: # 每训练十次，打印一次均方差
print("Epoch ", i, "MSE: ", loss.item())
hist[i] = loss.item()
#优化器优化模型
opitmizer.zero_grad()
loss.backward()
opitmizer.step()
y_train_pred=cnn_lstm(train_x)
loss_fn(y_train_pred,train_y).item()
y_test_pred=cnn_lstm(valid_x)
loss_fn(y_test_pred,valid_y)
plt.grid()
plt.xlabel("iters")
plt.ylabel("")
plt.title("loss", fontsize = 12)
plt.plot(hist, "r")
plt.show()
data_loader = valid_dataloader
# 存放测试序列的预测结果
predicts = []
# 存放测试序列的实际发生的结果
labels = []
for idx, (x, label) in enumerate(data_loader):
if (x.shape[0] != 64):
continue
# 对测试集样本进行批量预测，把结果保存到predict Tensor中
# 开环预测：即每一次序列预测与前后的序列无关。
predict= cnn_lstm(x)
# 把保存在tensor中的批量预测结果转换成list
predicts.extend(predict.data.squeeze(1).tolist())
# 把保存在tensor中的批量标签转换成list
labels.extend(label.data.squeeze(1).tolist())
predicts = np.array(predicts)
labels = np.array(labels)
print(predicts.shape)
print(labels.shape)
predicts_unnormalized = df_close_min + (df_close_max - df_close_min) * predicts
labels_unnormalized = df_close_min + (df_close_max - df_close_min) * labels
print("shape:", predicts_unnormalized.shape)
print("正则化后的预测数据：\n", predicts)
print("")
print("正则化前的预测数据：\n", predicts_unnormalized)
plt.plot(predicts_unnormalized,"r",label="pred")
plt.plot(labels_unnormalized, "b",label="real")
plt.show()
