Big Data-driven Electricity Plan Recommender System

摘要

The deregulation of electricity retailing market enables residential customers to select suitable electricity retailing plans to lower energy expenditures. This paper proposes a hybrid collaborative filtering-based electricity plan recommender system (HCF-EPRS), which is constructed in a two-stage model integrated with model-based and memory-based collaborative filtering skills. A weighted similarity metric is developed for a better similarity evaluation. Free from the requirements on whole house electricity usage data and historical plan transactions data, residential customers can obtain effective instructions in retailer and plan selection from HCF-EPRS through supplying some easily obtainable features. These features are the weekly operation duration times of some common household appliances. Through numerical tests on practical electricity users and retailing plans, HCF-EPRS is verified outperforming other approaches in recommending accuracy. Ideally, the instructions of HCF-EPRS on electricity retailer and plan selection helps to improve the competitive operation of the electricity market.