SINTYANINGRUM, NAYSILA (2017) PEMILIHAN INPUT DENGAN RANDOM FOREST PADA MODEL TIME SERIES REGRESSION DAN DOUBLE SEASONAL ARIMA UNTUK PERAMALAN DATA PEMAKAIAN BEBAN LISTRIK JANGKA PENDEK. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
1313100025_Undergraduate_Theses.pdf - Published Version Download (8MB) | Preview |
Abstract
Pemakaian beban listrik untuk menunjang aktivitas masyarakat semakin meningkat. Peningkatan permintaan energi listrik yang tinggi perlu diimbangi dengan ketersediaan pasokan listrik yang memadai. Ketersediaan energi listrik erat kaitannya dengan tinggi rendahnya daya listrik. Apabila pembangkit listrik menghasilkan daya yang terlalu besar maka akan terjadi pemborosan sedangkan jika daya yang dihasilkan terlalu kecil maka mengakibatkan pemadaman listrik secara bergilir. Berdasarkan hal tersebut, peramalan pemakaian beban listrik jangka pendek menjadi sangat penting. Model peramalan yang mempunyai akurasi tinggi sangat sulit didapatkan karena pemilihan variabel input pada model melibatkan lag yang sangat besar. Oleh karena itu dalam penelitian ini akan dibahas mengenai aplikasi metode Random Forest untuk pemilihan input (prediktor) pada model Time Series Regression dan Double Seasonal ARIMA untuk peramalan pemakaian beban listrik jangka pendek. Data yang digunakan merupakan data sekunder pemakaian beban listrik Jawa Timur per-setengah jam pada periode 24 September 2015 hingga 30 September 2016. Penelitian ini diperoleh kesimpulan bahwa peramalan terbaik menggunakan pemilihan input dengan metode Stepwise pada model Time Series Regression.
=============================================================================================================
The use of electrical load to support community activities is increasing. The enhancement of the demand for high electrical energy needs to be balanced with the availability of reliable electricity supply. Availability of electrical energy is closely related to the high and low power. If the power plant produces power that is too large then there would be a waste whereas if the power generated is too small, then it would lead to blackouts of the houses around. Accordingly, the use of forecasting short-term power load becomes very important. Forecasting model has a high accuracy is very difficult to obtain because of the selection of the input variables in the model involve a huge lag. Therefore, in this study there will be a discussion on application of Random Forest method for the selection of input (predictor) on the model Time Series Regression and Double Seasonal ARIMA for forecasting short-term electricity load consumption. The data used is secondary data usage East Java electrical loads per half-an-hour during the period 24 September 2015 to 30 September 2016. This study we concluded that the best forecasting using the input selection with Stepwise method on Time Series Regression model.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Uncontrolled Keywords: | Beban Listrik, Double Seasonal ARIMA, Time Series Regression, Random Forest, Electrical Load |
| Subjects: | H Social Sciences > HA Statistics |
| Divisions: | Faculty of Mathematics and Science > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | - NAYSILA SINTYANINGRUM |
| Date Deposited: | 26 Jan 2017 06:08 |
| Last Modified: | 08 Mar 2019 03:53 |
| URI: | http://repository.its.ac.id/id/eprint/3542 |
Actions (login required)
 |
View Item |