Interpretable LSTM Based on Mixture Attention Mechanism for Multi-Step Residential Load Forecasting
Open Access
- 13 July 2022
- journal article
- research article
- Published by MDPI AG in Electronics
- Vol. 11 (14) , 2189
Abstract
Residential load forecasting is of great significance to improve the energy efficiency of smart home services. Deep-learning techniques, i.e., long short-term memory (LSTM) neural networks, can considerably improve the performance of prediction models. However, these black-box networks are generally unexplainable, which creates an obstacle for the customer to deeply understand forecasting results and rapidly respond to uncertain circumstances, as practical engineering requires a high standard of prediction reliability. In this paper, an interpretable deep-learning method is proposed to solve the multi-step residential load forecasting problem which is referred to as explainable artificial intelligence (XAI). An encoder–decoder network architecture based on multi-variable LSTM (MV-LSTM) is developed for the multi-step probabilistic-load forecasting. The mixture attention mechanism is introduced in each prediction time step to better capture the different temporal dynamics of multivariate sequence in an interpretable form. By evaluating the contribution of each variable to the forecast, multi-quantile forecasts at multiple future time steps can be generated. The experiments on the real data set show that the proposed method can achieve good prediction performance while providing valuable explanations for the prediction results. The findings help end users gain insights into the forecasting model, bridging the gap between them and advanced deep-learning techniques.Keywords
Funding Information
- Fundamental Research Funds for the Central Universities of Central South University (2019zzts563, 2021JJ20082)
- Hunan Provincial Natural Science Foundation of China (2019zzts563, 2021JJ20082)
This publication has 29 references indexed in Scilit:
- Short-Term Residential Load Forecasting Based on LSTM Recurrent Neural NetworkIEEE Transactions on Smart Grid, 2017
- A Dual-Stage Attention-Based Recurrent Neural Network for Time Series Prediction2017
- Short-Term Residential Load Forecasting Based on Resident Behaviour LearningIEEE Transactions on Power Systems, 2017
- Deep Learning for Household Load Forecasting—A Novel Pooling Deep RNNIEEE Transactions on Smart Grid, 2017
- Energy-Sharing Model With Price-Based Demand Response for Microgrids of Peer-to-Peer ProsumersIEEE Transactions on Power Systems, 2017
- Probabilistic electric load forecasting: A tutorial reviewInternational Journal of Forecasting, 2016
- Clustering of Electricity Consumption Behavior Dynamics Toward Big Data ApplicationsIEEE Transactions on Smart Grid, 2016
- Photovoltaic and solar power forecasting for smart grid energy managementCSEE Journal of Power and Energy Systems, 2015
- Demand response and smart grids—A surveyRenewable and Sustainable Energy Reviews, 2014
- Impact of High PV Penetration on Voltage Profiles in Residential NeighborhoodsIEEE Transactions on Sustainable Energy, 2012