A New Optimization Technique Of Support Vector Machine For Electricity Market Price Forecasting
Forecasting price is an essential task in electrical power system. In terms of duration, it can be classified into three types, namely short term price forecasting, medium term price forecasting and long term price forecasting. Short term price forecasting such as day-ahead prediction provides forec...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Technical Report |
| Language: | en |
| Published: |
UTeM
2019
|
| Online Access: | http://eprints.utem.edu.my/id/eprint/25475/1/A%20New%20Optimization%20Technique%20Of%20Support%20Vector%20Machine%20For%20Electricity%20Market%20Price%20Forecasting.pdf http://eprints.utem.edu.my/id/eprint/25475/ https://plh.utem.edu.my/cgi-bin/koha/opac-detail.pl?biblionumber=118048 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Forecasting price is an essential task in electrical power system. In terms of duration, it can be classified into three types, namely short term price forecasting, medium term price forecasting and long term price forecasting. Short term price forecasting such as day-ahead prediction provides forecast prices for a day-ahead (24 hours) up to few days ahead that is useful for daily operation. Medium term forecast predicts for several weeks ahead up to a year or few months ahead. Meanwhile, long term price forecasting predicts yearly prices that is crucial for next year’s investment. The main challenge for electricity price forecasting is accuracy and efficiency of the forecast model. Lower accuracy is produced due to the nature of electricity price that is highly volatile. Hence, some researchers have developed complex procedures and techniques to produce more accurate forecast while considering significant feature selection as well as parameter optimization. On the other hand, some models ignore certain features that may contribute for forecast accuracy. Therefore, a hybrid of Least Square Support Vector Machine (LSSVM) and Bacterial Foraging Optimization Algorithms (BFOA) is proposed in this research to provide an accurate price forecast with optimized LSSVM parameters (gamma and sigma) and features. During optimization process, a huge number of features will be minimized and the LSSVM parameters will be optimized simultaneously. MATLAB will be used to simulate the performance of BFOA and evaluate the accuracy of LSSVM forecasting. BFOA has shown good performances in various field but it has yet to be attempted in electricity price forecasting. The forecast models are expected to provide higher accuracy than other existing models then examined on the same power market and test period. It can also be a mechanism for price forecasting in Malaysia when deregulated electricity market is being implemented by the government in the future. |
|---|