Performance of Deep Learning models with transfer learning for multiple-step-ahead forecasts in monthly time series

Authors

  • Martín Solís Tecnológico de Costa Rica, Costa Rica
  • Luis-Alexander Calvo-Valverde Tecnológico de Costa Rica, Costa Rica

DOI:

https://doi.org/10.4114/intartif.vol25iss70pp110-125

Keywords:

Deep Learning, Time Series, Transfer Learning, Machine Learning, Forescast

Abstract

Deep Learning and transfer learning models are being used to generate time series forecasts; however, there is scarce evidence about their performance prediction that it is more evident for monthly time series. The purpose of this paper is to compare Deep Learning models with transfer learning and without transfer learning and other traditional methods used for monthly forecasts to answer three questions about the suitability of Deep Learning and Transfer Learning to generate predictions of time series. Time series of M4 and M3 competitions were used for the experiments. The results suggest that deep learning models based on TCN, LSTM, and CNN with transfer learning tend to surpass the performance prediction of other traditional methods. On the other hand, TCN and LSTM, trained directly on the target time series, got similar or better performance than traditional methods for some forecast horizons.

Downloads

Download data is not yet available.

Metrics

Metrics Loading ...

Downloads

Published

2022-12-07

How to Cite

Solís, M., & Calvo-Valverde, L.-A. (2022). Performance of Deep Learning models with transfer learning for multiple-step-ahead forecasts in monthly time series. Inteligencia Artificial, 25(70), 110–125. https://doi.org/10.4114/intartif.vol25iss70pp110-125