An Explainable, Attention-Enhanced, Bidirectional Long Short-Term Memory Neural Network for Joint 48-Hour Forecasting of Temperature, Irradiance, and Relative Humidity

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Vamvouras, Georgios, Braimakis, Konstantinos, Tzivanidis, Christos
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915468628983808
author Vamvouras, Georgios
Braimakis, Konstantinos
Tzivanidis, Christos
author_facet Vamvouras, Georgios
Braimakis, Konstantinos
Tzivanidis, Christos
contents This paper presents a Deep Learning (DL) framework for 48-hour forecasting of temperature, solar irradiance, and relative humidity to support Model Predictive Control (MPC) in smart HVAC systems. The approach employs a stacked Bidirectional Long Short-Term Memory (BiLSTM) network with attention, capturing temporal and cross-feature dependencies by jointly predicting all three variables. Historical meteorological data (2019-2022) with encoded cyclical time features were used for training, while 2023 data evaluated generalization. The model achieved Mean Absolute Errors of 1.3 degrees Celsius (temperature), 31 W/m2 (irradiance), and 6.7 percentage points (humidity), outperforming state-of-the-art numerical weather prediction and machine learning benchmarks. Integrated Gradients quantified feature contributions, and attention weights revealed temporal patterns, enhancing interpretability. By combining multivariate forecasting, attention-based DL, and explainability, this work advances data-driven weather prediction. The demonstrated accuracy and transparency highlight the framework's potential for energy-efficient building control through reliable short-term meteorological forecasting.
format Preprint
id arxiv_https___arxiv_org_abs_2508_21109
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle An Explainable, Attention-Enhanced, Bidirectional Long Short-Term Memory Neural Network for Joint 48-Hour Forecasting of Temperature, Irradiance, and Relative Humidity
Vamvouras, Georgios
Braimakis, Konstantinos
Tzivanidis, Christos
Machine Learning
Artificial Intelligence
This paper presents a Deep Learning (DL) framework for 48-hour forecasting of temperature, solar irradiance, and relative humidity to support Model Predictive Control (MPC) in smart HVAC systems. The approach employs a stacked Bidirectional Long Short-Term Memory (BiLSTM) network with attention, capturing temporal and cross-feature dependencies by jointly predicting all three variables. Historical meteorological data (2019-2022) with encoded cyclical time features were used for training, while 2023 data evaluated generalization. The model achieved Mean Absolute Errors of 1.3 degrees Celsius (temperature), 31 W/m2 (irradiance), and 6.7 percentage points (humidity), outperforming state-of-the-art numerical weather prediction and machine learning benchmarks. Integrated Gradients quantified feature contributions, and attention weights revealed temporal patterns, enhancing interpretability. By combining multivariate forecasting, attention-based DL, and explainability, this work advances data-driven weather prediction. The demonstrated accuracy and transparency highlight the framework's potential for energy-efficient building control through reliable short-term meteorological forecasting.
title An Explainable, Attention-Enhanced, Bidirectional Long Short-Term Memory Neural Network for Joint 48-Hour Forecasting of Temperature, Irradiance, and Relative Humidity
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2508.21109