Comparative Performance Analysis of LSTM and Bidirectional LSTM for Short-Term Solar Power Generation Forecasting Using Limited Time-Series Data

DOI

10.2991/978-94-6239-731-6_5How to use a DOI?

Keywords

Bidirectional Long Short-Term Memory (BiLSTM); Long Short-Term Memory (LSTM); Photovoltaic; Renewable energy; Short-term forecasting

Abstract

Short-term forecasting of solar power generation plays a critical role in enhancing the reliability and efficiency of renewable energy integration into modern power systems. However, the inherent intermittency and variability of photovoltaic (PV) output present significant challenges, particularly under limited data conditions and in the absence of exogenous variables. This study aims to analyze and compare the performance of Long Short-Term Memory (LSTM) and Bidirectional Long Short-Term Memory (BiLSTM) models for short-term forecasting using a limited time-series dataset obtained from a residential PV system with a capacity of 2 × 500 Wp over a 15-day period. Model performance is evaluated using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination (R²). The results demonstrate that both models are capable of effectively capturing temporal dynamics, as indicated by R² values exceeding 0.81. Quantitatively, the LSTM model achieves an MAE of 50.11 W and an RMSE of 83.70 W, while the BiLSTM model shows slightly improved performance with an MAE of 48.12 W and an RMSE of 83.14 W. Nevertheless, the high MAPE values—187.33% for LSTM and 103.59% for BiLSTM—highlight the limitations of this metric when applied to data with near-zero actual values. Overall, BiLSTM provides a marginal performance improvement over LSTM; however, both models exhibit limitations in accurately predicting extreme values.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Muhamad Fahmi Amrillah
AU  - Bhakti Yudho Suprapto
AU  - Herlina Herlina
PY  - 2026
DA  - 2026/07/09
TI  - Comparative Performance Analysis of LSTM and Bidirectional LSTM for Short-Term Solar Power Generation Forecasting Using Limited Time-Series Data
BT  - Proceedings of the International Conference of Inland Water and Ferries Transport Polytechnic of Palembang on Technology and Environment (IWPOSPA-TE 2025)
PB  - Atlantis Press
SP  - 33
EP  - 40
SN  - 2352-5401
UR  - https://doi.org/10.2991/978-94-6239-731-6_5
DO  - 10.2991/978-94-6239-731-6_5
ID  - Amrillah2026
ER  -