机器学习预测

import matplotlib.pyplot as plt import pandas as pd from sklearn.linear_model import LinearRegression from sklearn.model_selection import train_test_split from sklearn.preprocessing import StandardScaler from sklearn.metrics import mean_squared_error as mse import matplotlib.pyplot as plt plt.rcParams['font.sans-serif']=['SimHei'] # 用来正常显示中文标签 plt.rcParams['axes.unicode_minus']=False # 用来正常显示负号 import seaborn as sns import pyecharts.options as opts from pyecharts.charts import Line data = pd.read_csv('黄金价格.csv') data = data.fillna(0) print(data.head(5)) # 设置时间为索引 data['Date'] = pd.to_datetime(data['Date']) # 重置时间顺序 data.set_index('Date', inplace=True) data.sort_values('Date', ascending=True, inplace=True) plt.figure(figsize=(12, 6)) data['Close/Last'].plot() plt.title("黄金价格走势图") plt.show() # 计算相关性系数 corr = data.corr() corr['Close/Last'].sort_values(ascending=False) ax = plt.subplots(figsize=(20, 16))#调整画布大小 ax = sns.heatmap(corr, vmax=.8, square=True, annot=True)#画热力图 annot=True 表示显示系数 # 设置刻度字体大小 plt.xticks(fontsize=20) plt.yticks(fontsize=20) plt.show() X = data.drop('Close/Last', axis=1) y = data['Close/Last'] # 数据标准化 scaler = StandardScaler() X = scaler.fit_transform(X) # 数据分割 X_train, X_test, y_train, y_test = train_test_split(X, y, shuffle=False,train_size=0.8,random_state=1) print(X_train) # 线性回归：--------- lr = LinearRegression() lr.fit(X_train, y_train) train_pred = lr.predict(X_train) test_pred = lr.predict(X_test) print(f'train MSE: {mse(y_train, train_pred):.5f}') print(f'test MSE: {mse(y_test, test_pred):.5f}') index = y_test.index plt.figure(figsize=(12, 6)) data['Close/Last'].plot(label='actual') pd.Series(test_pred, index=index).plot(color='red',label='pred') plt.title("线性回归预测结果对比图") plt.legend() plt.savefig("线性回归预测结果对比图.png") plt.show() # xgboost：--------- from xgboost import XGBRegressor lr = XGBRegressor() lr.fit(X_train, y_train) train_pred = lr.predict(X_train) test_pred = lr.predict(X_test) print(f'train MSE: {mse(y_train, train_pred):.5f}') print(f'test MSE: {mse(y_test, test_pred):.5f}') index = y_test.index plt.figure(figsize=(12, 6)) data['Close/Last'].plot(label='actual') pd.Series(test_pred, index=index).plot(color='red',label='pred') plt.title("xgboost预测结果对比图") plt.legend() plt.savefig("xgboost预测结果对比图.png") plt.show() from sklearn.preprocessing import MinMaxScaler from sklearn.model_selection import train_test_split import torch import torch.nn as nn import matplotlib.pyplot as plt df = pd.read_csv('黄金价格.csv') df = data.fillna(0) # 数据归一化 scaler = MinMaxScaler(feature_range=(-1, 1)) data = scaler.fit_transform(df['Close/Last'].values.reshape(-1, 1)) # 划分训练集和测试集 train_data, test_data = train_test_split(data, test_size=0.2, shuffle=False) # 转换为tensor train_data = torch.FloatTensor(train_data) test_data = torch.FloatTensor(test_data) # 定义LSTM模型 class LSTM(nn.Module): def __init__(self, input_size=1, hidden_layer_size=50, output_size=1): super().__init__() self.hidden_layer_size = hidden_layer_size self.lstm = nn.LSTM(input_size, hidden_layer_size) self.linear = nn.Linear(hidden_layer_size, output_size) self.hidden_cell = (torch.zeros(1,1,self.hidden_layer_size), torch.zeros(1,1,self.hidden_layer_size)) def forward(self, input_seq): lstm_out, self.hidden_cell = self.lstm(input_seq.view(len(input_seq) ,1, -1), self.hidden_cell) predictions = self.linear(lstm_out.view(len(input_seq), -1)) return predictions[-1] # 训练模型 model = LSTM() loss_function = nn.MSELoss() optimizer = torch.optim.Adam(model.parameters(), lr=0.001) epochs = 100 losses = [] for i in range(epochs): for seq in train_data: optimizer.zero_grad() model.hidden_cell = (torch.zeros(1, 1, model.hidden_layer_size), torch.zeros(1, 1, model.hidden_layer_size)) y_pred = model(seq) single_loss = loss_function(y_pred, seq) single_loss.backward() optimizer.step() losses.append(single_loss.item()) if i%25 == 1: print(f'epoch: {i:3} loss: {single_loss.item():10.8f}') print(f'epoch: {i:3} loss: {single_loss.item():10.10f}') import numpy as np # 使用训练好的模型进行预测 model.eval() test_inputs = train_data[-100:].tolist() test_outputs = [] for i in range(len(test_data)): seq = torch.FloatTensor(test_inputs[-100:]) with torch.no_grad(): model.hidden = (torch.zeros(1, 1, model.hidden_layer_size), torch.zeros(1, 1, model.hidden_layer_size)) test_inputs.append(test_data[i]) test_outputs.append(model(seq).item()) # 反归一化 predicted_price = scaler.inverse_transform(np.array(test_outputs).reshape(-1, 1)) # 绘制预测结果 plt.figure(figsize=(12, 6)) plt.plot(range(len(train_data), len(train_data) + len(predicted_price)), predicted_price, color='r', label='Predicted Price') plt.plot(range(len(data)), scaler.inverse_transform(data), color='b', label='Actual Price') plt.legend() plt.savefig("LSTM预测结果对比图.png") plt.show()