测试预测值后的Keras稳定下降并下降

Question

问题：

我正在尝试使用Keras中的LSTM模型预测Google的未来股价。我能够成功地训练模型，并且测试预测也进行得很好，但是后期测试/未来预测很差。它形成了一条稳定下降的曲线，这不是实际的未来数据。

一些说明

我正在用两个输入训练模型，并期望从中获得单个输出。

# Feature Scaling
from sklearn.preprocessing import MinMaxScaler
sc = MinMaxScaler(feature_range = (0, 1))
training_set_scaled = sc.fit_transform(training_set)

# Creating a data structure with 60 timesteps and 1 output
X_train = []
y_train = []
for i in range(2, 999):
    X_train.append(training_set_scaled[i-2:i, 0])
    y_train.append(training_set_scaled[i, 0])
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], 1))


# Part 2 - Building the RNN

# Importing the Keras libraries and packages
from keras.models import Sequential
from keras.layers import Dense
from keras.layers import LSTM
from keras.layers import Dropout

# Initialising the RNN
regressor = Sequential()

# Adding the first LSTM layer and some Dropout regularisation
regressor.add(LSTM(units = 50, return_sequences = True, input_shape = (X_train.shape[1], 1)))
regressor.add(Dropout(0.2))

# Adding a second LSTM layer and some Dropout regularisation
regressor.add(LSTM(units = 50, return_sequences = True))
regressor.add(Dropout(0.2))

# Adding a third LSTM layer and some Dropout regularisation
regressor.add(LSTM(units = 50, return_sequences = True))
regressor.add(Dropout(0.2))

# Adding a fourth LSTM layer and some Dropout regularisation
regressor.add(LSTM(units = 50))
regressor.add(Dropout(0.2))

# Adding the output layer
regressor.add(Dense(units = 1))

# Compiling the RNN
regressor.compile(optimizer = 'rmsprop', loss = 'mean_squared_error')

# Fitting the RNN to the Training set
regressor.fit(X_train, y_train, epochs = 500, batch_size = 50)

测试预测模型

dataset_test = pd.read_csv('/media/vinothubuntu/Ubuntu Storage/Downloads/Test - Test.csv')
real_stock_price = dataset_test.iloc[:, 2:3].values

# Getting the predicted stock price of 2017
dataset_total = pd.concat((dataset_train['data'], dataset_test['data']), axis = 0)
inputs = dataset_total[len(dataset_total) - len(dataset_test) -0:].values
inputs = inputs.reshape(-1,1)
inputs = sc.transform(inputs)
X_test = []
test_var = []
for i in range(0, 28):
    X_test.append(inputs[i:i+2, 0])
    test_var.append(inputs[i, 0])

X_test_pred = np.array(X_test)
X_test_pred = np.reshape(X_test_pred, (X_test_pred.shape[0], X_test_pred.shape[1], 1))
predicted_stock_price = regressor.predict(X_test_pred)

这部分进行得很好，测试预测给出了理想的结果。

测试/未来预测后：

for x in range(0,30):
    X_test_length = X_test[len(X_test)-1] # get the last array of X_test list   
    future=[]
    Prev_4 = X_test_length[1:2] # get the last four value of the X_test_length
    Last_pred = predicted_stock_price.flat[-1] # get the last value from prediction
    merger = np.append(Prev_4,Last_pred) 
    X_test.append(merger) #append the new array to X_test
    future.append(merger) #append the new array to future array 
    one_time_pred=np.array(future)
    one_time_pred = np.reshape(one_time_pred, (one_time_pred.shape[0], one_time_pred.shape[1], 1))   
    future_prediction = regressor.predict(one_time_pred) #predict future - gives one new prediction
    predicted_stock_price = np.append(predicted_stock_price, future_prediction, axis=0) #put the new predicction on predicted_stock_price array

这是实际的问题，我正在从测试预测中获取最后一个值，并预测单个输出，并在新的预设值上创建一个循环。 [如果您觉得这不是一个好主意，请提出一种更好的方法给我。]

我的输出：

预期结果：未来的实际数据，绝对不是下降曲线。