import numpy as np import pandas as pd import torch from torch import nn import os DAYS_FOR_TRAIN = 9 class LSTM_Regression(nn.Module): def __init__(self, input_size, hidden_size, output_size=1, num_layers=2): super().__init__() self.lstm = nn.LSTM(input_size, hidden_size, num_layers) self.fc = nn.Linear(hidden_size, output_size) def forward(self, _x): x, _ = self.lstm(_x) # _x is input, size (seq_len, batch, input_size) s, b, h = x.shape # x is output, size (seq_len, batch, hidden_size) x = x.view(s * b, h) x = self.fc(x) x = x.view(s, b, -1) # 把形状改回来 return x def create_dataset(data, days_for_train=5) -> (np.array, np.array): dataset_x, dataset_y = [], [] for i in range(len(data) - days_for_train): _x = data[i:(i + days_for_train)] dataset_x.append(_x) dataset_y.append(data[i + days_for_train]) return (np.array(dataset_x), np.array(dataset_y)) if __name__ == '__main__': data = pd.read_excel(r'C:\Users\鸽子\Desktop\浙江21年-23年行业以及分电压等级日电量.xlsx', sheet_name=1) data.columns = data.columns.map(lambda x: x.strip()) data.sort_values(by='pt_date', ascending=True) data['pt_date'] = pd.to_datetime(data['pt_date'], format='%Y%m%d', errors='coerce').astype('string') data.set_index('pt_date', inplace=True) data = data[(data.index.str[:6] != '202309') & (data.index.str[:6] != '202310')] data.drop(columns=[i for i in data.columns if (data[i] == 0).sum() / len(data) >= 0.5], inplace=True) # 去除0值列 print('len(data):', len(data)) for level in data.columns: df = data[level] df = df[df.values != 0] # 去除0值行 df = df.astype('float32').values # 转换数据类型 # 标准化到0~1 max_value = np.max(df) min_value = np.min(df) df = (df - min_value) / (max_value - min_value) dataset_x, dataset_y = create_dataset(df, DAYS_FOR_TRAIN) print('len(dataset_x:)', len(dataset_x)) # 划分训练集和测试集 train_size = len(dataset_x) - 31 train_x = dataset_x[:train_size] train_y = dataset_y[:train_size] # 将数据改变形状,RNN 读入的数据维度是 (seq_size, batch_size, feature_size) train_x = train_x.reshape(-1, 1, DAYS_FOR_TRAIN) train_y = train_y.reshape(-1, 1, 1) # 转为pytorch的tensor对象 train_x = torch.from_numpy(train_x) train_y = torch.from_numpy(train_y) model = LSTM_Regression(DAYS_FOR_TRAIN, 32, output_size=1, num_layers=2) # 导入模型并设置模型的参数输入输出层、隐藏层等 train_loss = [] loss_function = nn.MSELoss() optimizer = torch.optim.Adam(model.parameters(), lr=0.001, betas=(0.9, 0.999), eps=1e-08, weight_decay=0) for i in range(1200): out = model(train_x) loss = loss_function(out, train_y) loss.backward() optimizer.step() optimizer.zero_grad() train_loss.append(loss.item()) # 保存模型 # torch.save(model.state_dict(),save_filename) # torch.save(model.state_dict(),os.path.join(model_save_dir,model_file)) # for test model = model.eval() # 转换成测试模式 # model.load_state_dict(torch.load(os.path.join(model_save_dir,model_file))) # 读取参数 dataset_x = dataset_x.reshape(-1, 1, DAYS_FOR_TRAIN) # (seq_size, batch_size, feature_size) dataset_x = torch.from_numpy(dataset_x) pred_test = model(dataset_x) # 全量训练集 # 模型输出 (seq_size, batch_size, output_size) pred_test = pred_test.view(-1).data.numpy() pred_test = np.concatenate((np.zeros(DAYS_FOR_TRAIN), pred_test)) assert len(pred_test) == len(df) # 反归一化 pred_test = pred_test * (max_value - min_value) + min_value df = df * (max_value - min_value) + min_value pred_test = pred_test.reshape(-1) df = df.reshape(-1) # 打印指标 print(abs(pred_test[-31:] - df[-31:]).mean() / df[-31:].mean()) result_eight = pd.DataFrame({'pred_test': pred_test[-31:], 'real': df[-31:]}) target = (result_eight['pred_test'][-3:].sum() - result_eight['real'][-3:].sum()) / result_eight[ 'real'].sum() print(target) # with open(fr'C:\Users\鸽子\Desktop\各地市8月行业电量预测\{excel[:2]}.txt', 'a', encoding='utf-8') as f: # f.write(f'{excel[:2]}市{industry}行业月底偏差率:{round(target, 5)}\n')