用注意力增强的双向LSTM联合预测温湿光，精度高且可解释。
An Explainable, Attention-Enhanced, Bidirectional Long Short-Term Memory Neural Network for Joint 48-Hour Forecasting of Temperature, Irradiance, and Relative Humidity
- 双向LSTM+注意力机制捕捉时序与多变量依赖关系
- 48小时预测误差：温差1.3℃，辐照31W/m²，湿度6.7%点
- 支持智能建筑控温，适合需高透明度的能源系统
本文提出一种深度学习框架，用于支持智能暖通空调系统的模型预测控制（MPC），实现对温度、太阳辐射强度和相对湿度的48小时联合预测。该方法采用堆叠的双向长短期记忆（BiLSTM）网络结合注意力机制，通过同时预测三个变量来捕捉时间依赖性和跨特征关联。使用2019–2022年历史气象数据训练，含周期性时间编码特征；2023年数据用于评估泛化能力。模型在测试中达到均方误差：温度1.3℃，辐照强度31 W/m²，湿度6.7个百分点，优于当前最先进的数值天气预报和机器学习基准。利用集成梯度量化特征贡献，注意力权重揭示时间模式，提升可解释性。该框架融合多变量预测、注意力驱动的深度学习与可解释性，推动数据驱动的短时气象预测发展。其高精度与透明性表明其在实现节能建筑控制中的潜力。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。