my dataset is a univariate time series with one column as months, other column having demands for the corresponding months. My test dataset has NaN vals only for all the months. Can LSTM be used in this case and if so, how? I tried using LSTM but only got NaN values. If you could help me out, that'd be great. Thanks!