如何实现二进制分类的LSTM？

Question

我是深入学习的初学者。我试图实现二元分类的LSTM。我有11个特征（连续值）的EEG数据集和1个0或1的输出。受试者（人）正在观看一段5分钟长的视频，每30秒后，不管他们是否喜欢（1）它（0），他们都会给出他们的评论。这是一个时间序列数据，我认为LSTM或GRU可能会有所帮助。但是，这并没有给我一个好的结果。其中一个原因可能是选择了适当数量的层，每个层中没有神经元，也没有我用来预测下一个输出的先前数据点。我附上了我为此编写的代码。请告诉我有什么问题

import numpy as np
import pandas as pd
dataset_train = pd.read_csv('EEG_train.csv')
training_set_scaled = dataset_train.iloc[:, 0:12].values
X_train = []
y_train = []
# Creating a data structure with 60 timesteps and 1 output
for i in range(60, 882):
    X_train.append(training_set_scaled[i-60:i, 0:12])
    y_train.append(training_set_scaled[i, 11])
X_train, y_train = np.array(X_train), np.array(y_train)
# Reshaping
X_train = np.reshape(X_train, (X_train.shape[0], X_train.shape[1], 12))
from keras.models import Sequential
from keras.layers import Dense
from keras.layers import LSTM, GRU
from keras.layers import Dropout
clf = Sequential()
# Adding the first LSTM layer and some Dropout regularisation
clf.add(LSTM(units = 50, return_sequences = True, activation ='relu', input_shape = 
(X_train.shape[1], 12)))
clf.add(Dropout(0.2))

# Adding a second LSTM layer and some Dropout regularisation
clf.add(LSTM(units = 50, activation ='relu', return_sequences = False))
clf.add(Dropout(0.2))

# Adding the output layer
clf.add(Dense(units = 1, activation = 'sigmoid'))

# Compiling the RNN
clf.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])

# Fitting the RNN to the Training set
clf.fit(X_train, y_train, epochs = 100, batch_size = 128)
dataset_test = pd.read_csv('EEG_test.csv')
real_test_value = dataset_test.iloc[:, 11:12].values

# Getting the predicted
dataset_total = pd.concat((dataset_train, dataset_test), axis = 0)
inputs = dataset_total[len(dataset_total) - len(dataset_test) - 60:].values
inputs = inputs.reshape(-1,12)
X_test = []
for i in range(60, 439):
    X_test.append(inputs[i-60:i, 0:12])
X_test = np.array(X_test)
X_test = np.reshape(X_test, (X_test.shape[0], X_test.shape[1], 12))

# Predicting the result
from sklearn.metrics import accuracy_score,classification_report,confusion_matrix
predicted_test_value = regressor.predict(X_test)
accuracy=accuracy_score(real_test_value, predicted_test_value)

print("ACCURACY: "+str(accuracy))