| Autor: |
Ghatage, Nitin B., Patil, Pramod D., Shinde, Sagar |
| Předmět: |
|
| Zdroj: |
Journal Européen des Systèmes Automatisés; Dec2023, Vol. 56 Issue 6, p981-991, 11p |
| Abstrakt: |
Time-series forecasting is challenging in the real world. Both short-term and long-term forecasting are important in various fields of research and industry. Most forecasting algorithms perform great in providing one-step predictions, i.e., predicting only the next value in the time series data, but do not perform well while predicting multiple steps into the future. On top of that, concept drift makes it more challenging. The aim of this paper is to develop a lightweight recurrent neural networks (RNN)-based model that can do forecasting in the short or long term with the ability to detect concept drift and adapt to it automatically using a recent window of the data stream. The suggested model performs better than current techniques, with the lowest Root Mean Square Error (RMSE) of 0.0701, demonstrating increased accuracy in adaptive time series forecasting for temperature control in smart homes. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Supplemental Index |
| Externí odkaz: |
|
