High-order adaptive model to forecast regional electricity loads

Over the past few years, a considerable number of studies have been proposed on load forecasting. This paper aims at proposing a promising model using high-order adaptive fuzzy time-series algorithm to get more efficient forecasting. From the reviewed literature related to fuzzy time-series, there are two points need to be concerned. The first is to determine a reasonable universe of discourse and the length of intervals, and the second is many researchers ignore the information of trend patterns change in the past history. Hence, this paper utilized the trend weighted and high order adaptive model to deal with above drawbacks. The proposed model is applied for forecasting the regional electricity load in Taiwan. The experiment results showed that the proposed model outperforms the listing methods under MAPE (mean absolute percentage error) criteria.