Weighted subsethood segmented fuzzy time series for moving holiday electricity load demand forecasting
Moving holiday is a non-fixed holiday according to the Gregorian calendar. Most of the electricity load demand studies showed that this event affects the accuracy of load forecasting. It is due to limited recent historical data about moving holiday, a longer time series is acquired to reveal the pat...
| Main Authors: | Mansor, R., Kasim, M.M., Othman, M. |
|---|---|
| Format: | Article |
| Institution: | Universiti Teknologi Petronas |
| Record Id / ISBN-0: | utp-eprints.23188 / |
| Published: |
Institute of Advanced Scientific Research, Inc.
2020
|
| Online Access: |
https://www.scopus.com/inward/record.uri?eid=2-s2.0-85084356875&doi=10.5373%2fJARDCS%2fV12SP2%2fSP20201122&partnerID=40&md5=74edeaee9c1dae6fd8c1f09c7282ab35 http://eprints.utp.edu.my/23188/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: |
Moving holiday is a non-fixed holiday according to the Gregorian calendar. Most of the electricity load demand studies showed that this event affects the accuracy of load forecasting. It is due to limited recent historical data about moving holiday, a longer time series is acquired to reveal the pattern, different characteristics of each moving holiday and existence of great amount of irregularities in the load data. All these matters contributed to the uncertainty in load demand forecasting. This paper modifies the classical fuzzy time series (FTS) algorithm by applying weighted subsethood based algorithm (WSBA) in FTS algorithm using segmented electricity load demand time series data. The modified algorithm, Weighted Subsethood Segmented Fuzzy Time Series (WeSuSFTS) consists of three main phases; data pre-processing, forecasting based on WeSuSFTS model and model evaluation. Step by step explanation for each phase is also presented. The results show that the WeSuSFTS algorithm can be one alternative electricity moving holiday load demand forecasting method and the algorithm achieves its lowest mean absolute percentage error (MAPE) at 2.8 only. © 2020, Institute of Advanced Scientific Research, Inc. All rights reserved. |
|---|