PERAMALAN BEBAN PUNCAK PEMAKAIAN LISTRIK DI AREA SEMARANG DENGAN METODE HYBRID ARIMA (AUTOREGRESSIVE INTEGRATED MOVING AVERAGE) – ANFIS (ADAPTIVE NEURO FUZZY INFERENCE SYSTEM) (Studi Kasus di PT PLN (Persero) Distribusi Jawa Tengah dan DIY)

KRISTIANA, ANA (2015) PERAMALAN BEBAN PUNCAK PEMAKAIAN LISTRIK DI AREA SEMARANG DENGAN METODE HYBRID ARIMA (AUTOREGRESSIVE INTEGRATED MOVING AVERAGE) – ANFIS (ADAPTIVE NEURO FUZZY INFERENCE SYSTEM) (Studi Kasus di PT PLN (Persero) Distribusi Jawa Tengah dan DIY). Undergraduate thesis, FSM Universitas Diponegoro.

[img]
Preview
PDF
2812Kb

Abstract

Electricity become one of the basic needs in society, so that the demand level for electricity even bigger as more complex activities in society. In order to fulfill the needs of electricity in Indonesia, PT PLN have to do electrical peak load forecasting to prevent electrical crisis. In this research, we use hybrid ARIMA-ANFIS methods to forecast daily peak load of electricity in Semarang period December 2014 until January 2015. The use of hybrid ARIMA-ANFIS is to capture both linear and nonlinear patterns in the data, because sometimes time series data can contain both linear and nonlinear patterns. Since ARIMA can not deal with nonlinear patterns while ANFIS is not able to handle both linear and nonlinear patterns alone. The accuracy of the model was measured by symmetric MAPE (sMAPE) criteria, in which the best model chosen is the model with the smallest sMAPE value. The results showed that the hybrid ARIMA-ANFIS model that used to predict the daily peak load electricity in Semarang during the period of December 2014 until January 2015, comes from combination between SARIMA (0,1,1)(0,1,1)7 model and residual forecasting with ANFIS model using first lag input, Gaussian membership function in 3 clusters. Keywords: Electricity, Electrical peak load forecasting, ARIMA, ANFIS, Hybrid ARIMA-ANFIS

Item Type:Thesis (Undergraduate)
Subjects:H Social Sciences > HA Statistics
Divisions:Faculty of Science and Mathematics > Department of Statistics
ID Code:47271
Deposited By:INVALID USER
Deposited On:05 Jan 2016 08:44
Last Modified:05 Jan 2016 08:44

Repository Staff Only: item control page