An integrated method based on relevance vector machine for short-term load forecasting

•A relevance vector machine (RVM) based integrated method for load demand estimation is proposed.•Features extracted are divided into two categories, current state related features and historical information related features.•The method is applicable to one-hour-ahead load forecasting and day-ahead load profile forecasting.•Detailed empirical comparison is carried out among the proposed method, time series methods, and artificial intelligence methods.•The proposed method is proven to have the highest precision.