TCN architecture for time series
Temporal convolutional network was evaluated with respect to following points:
- Split selected dataset into training, evaluation and testing subsets
- Test set has lenght of
seq_length + n_ahead
whereseq_length
is length of sequence fed to TCN, andn_ahead
is prediction length - Train network using training and evaluation sets
- Feed
seq_lenght
of test set into trained network and getn_ahead
predictions - Evaluate
n_ahead
predictions againstn_ahead
from test set
- compare TCN to other models (LSTM, Holt Winters)