Explainable AI for predicting daily household energy usages

In the recent era, for most sustainable smart cities energy conservation is a major point of consideration as urbanization is been carried out at an exponential rate. Out of that most of the energy consumption is diverted toward households, and there is a huge scope for optimization of this energy. Hence predicting this household energy with the advancement of AI and Machine Learning techniques is considered a social contribution and area of interest for most researchers. But only predicting the energy consumption will not solve the problem of energy optimization for a city, it is also important to understand the factors responsible for such predictions so that all possible recourses could be carried out to those factors, and it becomes more accountable, trustworthy and justifiable its energy optimization decisions towards its all stakeholders. With the use of Explainable Artificial Intelligence (XAI) techniques such as LIME, SHAP is possible to improve the explainability of machine learning models. Here SHAP technique is adopted to understand the prediction model and identify mostly responsible factors for the energy consumption of households.