快捷導(dǎo)航

PyTorch搭建LSTM實(shí)現(xiàn)時(shí)間序列負(fù)荷預(yù)測(cè)

更新時(shí)間：2022年05月11日 10:06:47 作者：Cyril_KI

這篇文章主要為大家介紹了PyTorch搭建LSTM實(shí)現(xiàn)時(shí)間序列負(fù)荷預(yù)測(cè)，有需要的朋友可以借鑒參考下，希望能夠有所幫助，祝大家多多進(jìn)步，早日升職加薪

I. 前言

在上一篇文章深入理解PyTorch中LSTM的輸入和輸出（從input輸入到Linear輸出）中，我詳細(xì)地解釋了如何利用PyTorch來(lái)搭建一個(gè)LSTM模型，本篇文章的主要目的是搭建一個(gè)LSTM模型用于時(shí)間序列預(yù)測(cè)。

系列文章：

PyTorch搭建LSTM實(shí)現(xiàn)多變量多步長(zhǎng)時(shí)序負(fù)荷預(yù)測(cè)

PyTorch搭建LSTM實(shí)現(xiàn)多變量時(shí)序負(fù)荷預(yù)測(cè)

PyTorch深度學(xué)習(xí)LSTM從input輸入到Linear輸出

PyTorch搭建雙向LSTM實(shí)現(xiàn)時(shí)間序列負(fù)荷預(yù)測(cè)

II. 數(shù)據(jù)處理

數(shù)據(jù)集為某個(gè)地區(qū)某段時(shí)間內(nèi)的電力負(fù)荷數(shù)據(jù)，除了負(fù)荷以外，還包括溫度、濕度等信息。

本篇文章暫時(shí)不考慮其它變量，只考慮用歷史負(fù)荷來(lái)預(yù)測(cè)未來(lái)負(fù)荷。

本文中，我們根據(jù)前24個(gè)時(shí)刻的負(fù)荷下一時(shí)刻的負(fù)荷。有關(guān)多變量預(yù)測(cè)請(qǐng)參考：PyTorch搭建LSTM實(shí)現(xiàn)多變量時(shí)間序列預(yù)測(cè)（負(fù)荷預(yù)測(cè)）。

def load_data(file_name):
    global MAX, MIN
    df = pd.read_csv('data/new_data/' + file_name, encoding='gbk')
    columns = df.columns
    df.fillna(df.mean(), inplace=True)
    MAX = np.max(df[columns[1]])
    MIN = np.min(df[columns[1]])
    df[columns[1]] = (df[columns[1]] - MIN) / (MAX - MIN)
    return df
class MyDataset(Dataset):
    def __init__(self, data):
        self.data = data
    def __getitem__(self, item):
        return self.data[item]
    def __len__(self):
        return len(self.data)
def nn_seq(file_name, B):
    print('處理數(shù)據(jù):')
    data = load_data(file_name)
    load = data[data.columns[1]]
    load = load.tolist()
    load = torch.FloatTensor(load).view(-1)
    data = data.values.tolist()
    seq = []
    for i in range(len(data) - 24):
        train_seq = []
        train_label = []
        for j in range(i, i + 24):
            train_seq.append(load[j])
        train_label.append(load[i + 24])
        train_seq = torch.FloatTensor(train_seq).view(-1)
        train_label = torch.FloatTensor(train_label).view(-1)
        seq.append((train_seq, train_label))
    # print(seq[:5])
    Dtr = seq[0:int(len(seq) * 0.7)]
    Dte = seq[int(len(seq) * 0.7):len(seq)]
    train_len = int(len(Dtr) / B) * B
    test_len = int(len(Dte) / B) * B
    Dtr, Dte = Dtr[:train_len], Dte[:test_len]
    train = MyDataset(Dtr)
    test = MyDataset(Dte)
    Dtr = DataLoader(dataset=train, batch_size=B, shuffle=False, num_workers=0)
    Dte = DataLoader(dataset=test, batch_size=B, shuffle=False, num_workers=0)
    return Dtr, Dte

上面代碼用了DataLoader來(lái)對(duì)原始數(shù)據(jù)進(jìn)行處理，最終得到了batch_size=B的數(shù)據(jù)集Dtr和Dte，Dtr為訓(xùn)練集，Dte為測(cè)試集。

III. LSTM模型

這里采用了深入理解PyTorch中LSTM的輸入和輸出（從input輸入到Linear輸出）中的模型：

class LSTM(nn.Module):
    def __init__(self, input_size, hidden_size, num_layers, output_size, batch_size):
        super().__init__()
        self.input_size = input_size
        self.hidden_size = hidden_size
        self.num_layers = num_layers
        self.output_size = output_size
        self.num_directions = 1 # 單向LSTM
        self.batch_size = batch_size
        self.lstm = nn.LSTM(self.input_size, self.hidden_size, self.num_layers, batch_first=True)
        self.linear = nn.Linear(self.hidden_size, self.output_size)
    def forward(self, input_seq):
        h_0 = torch.randn(self.num_directions * self.num_layers, self.batch_size, self.hidden_size).to(device)
        c_0 = torch.randn(self.num_directions * self.num_layers, self.batch_size, self.hidden_size).to(device)
        seq_len = input_seq.shape[1] # (5, 24)
        # input(batch_size, seq_len, input_size)
        input_seq = input_seq.view(self.batch_size, seq_len, 1)  # (5, 24, 1)
        # output(batch_size, seq_len, num_directions * hidden_size)
        output, _ = self.lstm(input_seq, (h_0, c_0)) # output(5, 24, 64)
        output = output.contiguous().view(self.batch_size * seq_len, self.hidden_size) # (5 * 24, 64)
        pred = self.linear(output) # pred(150, 1)
        pred = pred.view(self.batch_size, seq_len, -1) # (5, 24, 1)
        pred = pred[:, -1, :]  # (5, 1)
        return pred

IV. 訓(xùn)練

def LSTM_train(name, b):
    Dtr, Dte = nn_seq(file_name=name, B=b)
    input_size, hidden_size, num_layers, output_size = 1, 64, 5, 1
    model = LSTM(input_size, hidden_size, num_layers, output_size, batch_size=b).to(device)
    loss_function = nn.MSELoss().to(device)
    optimizer = torch.optim.Adam(model.parameters(), lr=0.001)
    # 訓(xùn)練
    epochs = 15
    cnt = 0
    for i in range(epochs):
        cnt = 0
        print('當(dāng)前', i)
        for (seq, label) in Dtr:
            cnt += 1
            seq = seq.to(device)
            label = label.to(device)
            y_pred = model(seq)
            loss = loss_function(y_pred, label)
            optimizer.zero_grad()
            loss.backward()
            optimizer.step()
            if cnt % 100 == 0:
                print('epoch', i, ':', cnt - 100, '~', cnt, loss.item())
    state = {'model': model.state_dict(), 'optimizer': optimizer.state_dict()}
    torch.save(state, LSTM_PATH)

一共訓(xùn)練了15輪：

V. 測(cè)試

def test(name, b):
    global MAX, MIN
    Dtr, Dte = nn_seq(file_name=name, B=b)
    pred = []
    y = []
    print('loading model...')
    input_size, hidden_size, num_layers, output_size = 1, 64, 5, 1
    model = LSTM(input_size, hidden_size, num_layers, output_size, batch_size=b).to(device)
    model.load_state_dict(torch.load(LSTM_PATH)['model'])
    model.eval()
    print('predicting...')
    for (seq, target) in Dte:
        target = list(chain.from_iterable(target.data.tolist()))
        y.extend(target)
        seq = seq.to(device)
        seq_len = seq.shape[1]
        seq = seq.view(model.batch_size, seq_len, 1)  # (5, 24, 1)
        with torch.no_grad():
            y_pred = model(seq)
            y_pred = list(chain.from_iterable(y_pred.data.tolist()))
            pred.extend(y_pred)
    y, pred = np.array(y), np.array(pred)
    y = (MAX - MIN) * y + MIN
    pred = (MAX - MIN) * pred + MIN
    print('accuracy:', get_mape(y, pred))
    # plot
    x = [i for i in range(1, 151)]
    x_smooth = np.linspace(np.min(x), np.max(x), 600)
    y_smooth = make_interp_spline(x, y[0:150])(x_smooth)
    plt.plot(x_smooth, y_smooth, c='green', marker='*', ms=1, alpha=0.75, label='true')
    y_smooth = make_interp_spline(x, pred[0:150])(x_smooth)
    plt.plot(x_smooth, y_smooth, c='red', marker='o', ms=1, alpha=0.75, label='pred')
    plt.grid(axis='y')
    plt.legend()
    plt.show()

MAPE為6.07%：

VI. 源碼及數(shù)據(jù)

源碼及數(shù)據(jù)我放在了GitHub上，LSTM-Load-Forecasting

以上就是PyTorch搭建LSTM實(shí)現(xiàn)時(shí)間序列負(fù)荷預(yù)測(cè)的詳細(xì)內(nèi)容，更多關(guān)于PyTorch搭建LSTM時(shí)間序列負(fù)荷預(yù)測(cè)的資料請(qǐng)關(guān)注腳本之家其它相關(guān)文章！

您可能感興趣的文章: