Correct selection of the y_train predictor

[IgnatiusDean]

Hi,

I was reviewing your 012 Stock Market Prediction using Multivariate Time Series Models and Python.ipynb code as it does everything I am looking for (Time Series, Train/Test Split, and actual forecasts).

In Section 3 Preprocessing and Feature Selection, when splitting the train data set into the X Features and Y Outcome/Predictors, it appears to me that you are assigning the first column of the DF to the Y train data set.

# The RNN needs data with the format of [samples, time steps, features].
# Here, we create N samples, 100 time steps per sample, and 2 features
for i in range(100, train_data_len):
    x_train.append(train_data[i-sequence_length:i,:]) #contains 100 values 0-100 * columsn
    y_train.append(train_data[i, 0]) #contains the prediction values for validation

When I execute this code it appears to me that 'column 0' in the train_data is the 'High' price, not the closing or opening prioe. Ultimately predicting the highest sale/purchase price for the next day... Is this a correct interpretation?

If so, if we wanted to predict the closing price, we could use

# The RNN needs data with the format of [samples, time steps, features].
# Here, we create N samples, 100 time steps per sample, and 2 features
for i in range(100, train_data_len):
    x_train.append(train_data[i-sequence_length:i,:]) #contains 100 values 0-100 * columsn
    y_train.append(train_data[i, 3]) #contains the prediction values for validation

Looking forwar