N-Predictions

[BehnamTaki]

Hi!
I read "Long Short Term Memory Networks for Anomaly Detection in Time Series" article. Their prediction model learns to predict the next 'm' values. so it means each time-step(every point of time series) has 'm' error prediction values and the output would be a t*m sequence, 't' depicts the length of input. am I right? in your code the N-prediction file is a 1-D sequence.what's different? and why?

Bests