Household energy consumption prediction by feature selection of lifestyle data

This study proposes a prediction model for the annual power consumption of general households based on the large-scale questionnaire about lifestyle of the household. By combining a feature selection technique with the multivariate linear regression model, we can find important features, i.e., key questionnaires, for estimating seasonal electricity consumption for ordinary households. This may be the first study on feature selection for actual questionnaire data in the energy field. Through numerical experiments, we confirmed the high prediction performance of our method by comparing the test accuracy of our method with that of a naive one.