@article{ART003130853},
author={Jaehyeok Jo and Yunho Sin and Bo-Young Kim and Jihoon Moon},
title={Comparative Analysis of RNN Architectures and Activation Functions with Attention Mechanisms for Mars Weather Prediction},
journal={Journal of The Korea Society of Computer and Information},
issn={1598-849X},
year={2024},
volume={29},
number={10},
pages={1-9}
TY - JOUR
AU - Jaehyeok Jo
AU - Yunho Sin
AU - Bo-Young Kim
AU - Jihoon Moon
TI - Comparative Analysis of RNN Architectures and Activation Functions with Attention Mechanisms for Mars Weather Prediction
JO - Journal of The Korea Society of Computer and Information
PY - 2024
VL - 29
IS - 10
PB - The Korean Society Of Computer And Information
SP - 1
EP - 9
SN - 1598-849X
AB - In this paper, we propose a comparative analysis to evaluate the impact of activation functions and attention mechanisms on the performance of time-series models for Mars meteorological data. Mars meteorological data are nonlinear and irregular due to low atmospheric density, rapid temperature variations, and complex terrain. We use long short-term memory (LSTM), bidirectional LSTM (BiLSTM), gated recurrent unit (GRU), and bidirectional GRU (BiGRU) architectures to evaluate the effectiveness of different activation functions and attention mechanisms. The activation functions tested include rectified linear unit (ReLU), leaky ReLU, exponential linear unit (ELU), Gaussian error linear unit (GELU), Swish, and scaled ELU (SELU), and model performance was measured using mean absolute error (MAE) and root mean square error (RMSE) metrics. Our results show that the integration of attentional mechanisms improves both MAE and RMSE, with Swish and ReLU achieving the best performance for minimum temperature prediction. Conversely, GELU and ELU were less effective for pressure prediction. These results highlight the critical role of selecting appropriate activation functions and attention mechanisms in improving model accuracy for complex time-series forecasting.
KW - Mars Weather Prediction;Time-Series Analysis;Attention Mechanism;Activation Functions;Deep Learning;Predictive Modeling
DO -
UR -
ER -
Jaehyeok Jo, Yunho Sin, Bo-Young Kim and Jihoon Moon. (2024). Comparative Analysis of RNN Architectures and Activation Functions with Attention Mechanisms for Mars Weather Prediction. Journal of The Korea Society of Computer and Information, 29(10), 1-9.
Jaehyeok Jo, Yunho Sin, Bo-Young Kim and Jihoon Moon. 2024, "Comparative Analysis of RNN Architectures and Activation Functions with Attention Mechanisms for Mars Weather Prediction", Journal of The Korea Society of Computer and Information, vol.29, no.10 pp.1-9.
Jaehyeok Jo, Yunho Sin, Bo-Young Kim, Jihoon Moon "Comparative Analysis of RNN Architectures and Activation Functions with Attention Mechanisms for Mars Weather Prediction" Journal of The Korea Society of Computer and Information 29.10 pp.1-9 (2024) : 1.
Jaehyeok Jo, Yunho Sin, Bo-Young Kim, Jihoon Moon. Comparative Analysis of RNN Architectures and Activation Functions with Attention Mechanisms for Mars Weather Prediction. 2024; 29(10), 1-9.
Jaehyeok Jo, Yunho Sin, Bo-Young Kim and Jihoon Moon. "Comparative Analysis of RNN Architectures and Activation Functions with Attention Mechanisms for Mars Weather Prediction" Journal of The Korea Society of Computer and Information 29, no.10 (2024) : 1-9.
Jaehyeok Jo; Yunho Sin; Bo-Young Kim; Jihoon Moon. Comparative Analysis of RNN Architectures and Activation Functions with Attention Mechanisms for Mars Weather Prediction. Journal of The Korea Society of Computer and Information, 29(10), 1-9.
Jaehyeok Jo; Yunho Sin; Bo-Young Kim; Jihoon Moon. Comparative Analysis of RNN Architectures and Activation Functions with Attention Mechanisms for Mars Weather Prediction. Journal of The Korea Society of Computer and Information. 2024; 29(10) 1-9.
Jaehyeok Jo, Yunho Sin, Bo-Young Kim, Jihoon Moon. Comparative Analysis of RNN Architectures and Activation Functions with Attention Mechanisms for Mars Weather Prediction. 2024; 29(10), 1-9.
Jaehyeok Jo, Yunho Sin, Bo-Young Kim and Jihoon Moon. "Comparative Analysis of RNN Architectures and Activation Functions with Attention Mechanisms for Mars Weather Prediction" Journal of The Korea Society of Computer and Information 29, no.10 (2024) : 1-9.
