利用 RNN 对 ECG 心电信号进行心拍分类（pytorch）

'''
导入相关包
'''
import wfdb
import pywt
import seaborn
import numpy as np
import matplotlib.pyplot as plt
from sklearn.metrics import confusion_matrix
import torch
import torch.utils.data as Data
from torch import nn

'''
加载数据集
'''

# 测试集在数据集中所占的比例
RATIO = 0.2

# 小波去噪预处理
def denoise(data):
    # 小波变换
    coeffs = pywt.wavedec(data=data, wavelet='db5', level=9)
    cA9, cD9, cD8, cD7, cD6, cD5, cD4, cD3, cD2, cD1 = coeffs
    # 阈值去噪
    threshold = (np.median(np.abs(cD1)) / 0.6745) * (np.sqrt(2 * np.log(len(cD1))))
    cD1.fill(0)
    cD2.fill(0)
    for i in range(1, len(coeffs) - 2):
        coeffs[i] = pywt.threshold(coeffs[i], threshold)
    # 小波反变换,获取去噪后的信号
    rdata = pywt.waverec(coeffs=coeffs, wavelet='db5')
    return rdata

# 读取心电数据和对应标签,并对数据进行小波去噪
def getDataSet(number, X_data, Y_data):
    ecgClassSet = ['N', 'A', 'V', 'L', 'R']
    # 读取心电数据记录
    # print("正在读取 " + number + " 号心电数据...")
    # 读取MLII导联的数据
    record = wfdb.rdrecord('C:/mycode/dataset_make/mit-bih-arrhythmia-database-1.0.0/' + number, channel_names=['MLII'])
    data = record.p_signal.flatten()
    rdata = denoise(data=data)
    # 获取心电数据记录中R波的位置和对应的标签
    annotation = wfdb.rdann('C:/mycode/dataset_make/mit-bih-arrhythmia-database-1.0.0/' + number, 'atr')
    Rlocation = annotation.sample
    Rclass = annotation.symbol
    # 去掉前后的不稳定数据
    start = 10
    end = 5
    i = start
    j = len(annotation.symbol) - end
    # 因为只选择NAVLR五种心电类型,所以要选出该条记录中所需要的那些带有特定标签的数据,舍弃其余标签的点
    # X_data在R波前后截取长度为300的数据点
    # Y_data将NAVLR按顺序转换为01234
    while i < j:
        try:
            # Rclass[i] 是标签
            lable = ecgClassSet.index(Rclass[i])  # 这一步就是相当于抛弃了不在ecgClassSet的索引
            # 基于经验值，基于R峰向前取100个点，向后取200个点
            x_train = rdata[Rlocation[i] - 100:Rlocation[i] + 200]
            X_data.append(x_train)
            Y_data.append(lable)
            i += 1
        except ValueError:
            i += 1
    return

# 加载数据集并进行预处理
def loadData():
    numberSet = ['100', '101', '103', '105', '106', '107', '108', '109', '111', '112', '113', '114', '115',
                 '116', '117', '119', '121', '122', '123', '124', '200', '201', '202', '203', '205', '208',
                 '210', '212', '213', '214', '215', '217', '219', '220', '221', '222', '223', '228', '230',
                 '231', '232', '233', '234']
    dataSet = []
    lableSet = []
    for n in numberSet:
        getDataSet(n, dataSet, lableSet)
    # 转numpy数组,打乱顺序
    dataSet = np.array(dataSet).reshape(-1, 300)  # 转化为二维，一行有 300 个，行数需要计算
    lableSet = np.array(lableSet).reshape(-1, 1)  # 转化为二维，一行有   1 个，行数需要计算
    train_ds = np.hstack((dataSet, lableSet))  # 将数据集和标签集水平堆叠在一起,(92192, 300) (92192, 1) (92192, 301)
    # print(dataSet.shape, lableSet.shape, train_ds.shape)  # (92192, 300) (92192, 1) (92192, 301)
    np.random.shuffle(train_ds)
    # 数据集及其标签集
    X = train_ds[:, :300].reshape(-1, 1, 300)  # (92192, 1, 300)
    Y = train_ds[:, 300]  # (92192)

    # 测试集及其标签集
    shuffle_index = np.random.permutation(len(X))  # 生成0-(X-1)的随机索引数组

    # 设定测试集的大小 RATIO是测试集在数据集中所占的比例
    test_length = int(RATIO * len(shuffle_index))
    # 测试集的长度
    test_index = shuffle_index[:test_length]
    # 训练集的长度
    train_index = shuffle_index[test_length:]
    X_test, Y_test = X[test_index], Y[test_index]
    X_train, Y_train = X[train_index], Y[train_index]
    return X_train, Y_train, X_test, Y_test

X_train, Y_train, X_test, Y_test = loadData()

'''
数据处理
'''
train_Data = Data.TensorDataset(torch.Tensor(X_train), torch.Tensor(Y_train)) # 返回结果为一个个元组，每一个元组存放数据和标签
train_loader = Data.DataLoader(dataset=train_Data, batch_size=128)
test_Data = Data.TensorDataset(torch.Tensor(X_test), torch.Tensor(Y_test)) # 返回结果为一个个元组，每一个元组存放数据和标签
test_loader = Data.DataLoader(dataset=test_Data, batch_size=128)


'''
模型搭建
'''
class RnnModel(nn.Module):
    def __init__(self):
        super(RnnModel, self).__init__()
        '''
        参数解释：(输入维度，隐藏层维度，网络层数)
        '''
        self.rnn = nn.RNN(300, 50, 3, nonlinearity='tanh')
        self.linear = nn.Linear(50, 5)

    def forward(self, x):
        r_out, h_state = self.rnn(x)
        output = self.linear(r_out[:,-1,:])  # 将 RNN 层的输出 r_out 在最后一个时间步上的输出（隐藏状态）传递给线性层
        return output

model = RnnModel()

'''
设置损失函数和参数优化方法
'''
criterion = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(model.parameters(), lr=0.01)

'''
模型训练
'''
EPOCHS = 5
for epoch in range(EPOCHS):
    running_loss = 0
    for i, data in enumerate(train_loader):
        inputs, label = data
        y_predict = model(inputs)
        loss = criterion(y_predict, label.long())
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()
        running_loss += loss.item()

    # 预测
    correct = 0
    total = 0
    with torch.no_grad():
        for data in test_loader:
            inputs, label = data
            y_pred = model(inputs)
            _, predicted = torch.max(y_pred.data, dim=1)
            total += label.size(0)
            correct += (predicted == label).sum().item()

    print(f'Epoch: {epoch + 1}, ACC on test: {correct / total}')

Code