<![CDATA[In system electricity, strategy predictions will need electricity very needed for anticipate need energy electricity in Indonesia, especially in East Java. Because East Java is one of them province with economy and growth high population facing challenge big in fulfil need the electricity in a way sustainable. Therefore, it is necessary predictions accurate for determine need electricity in distribution electricity period long that is annual. Prediction burden electricity in the area East Java uses Adaptive Neuro Fuzzy Inference System (ANFIS) method. Method ANFIS research through stages that is studies literature, collection, data processing, data simulated in Matlab with input training and testing data, ANFIS training, ANFIS testing, and analysis results and conclusions. In this prediction, 2 scenarios were carried out. Research result shows: 1) Scenario 1 has 3 inputs, including quantity population(t), GRDP (t), and burden electricity (t), as well as 1 target output, namely burden electricity (t+1). Scenario 2 has 4 inputs, including amount population (t), GRDP (t), burden electricity (t-1), and load electricity (t), as well as 1 target output, namely (t+1). This research uses a Generalized Bell type membership function with 3 membership functions for each input data with the number of epochs is 100 times; 2) From 2 scenarios the produces the highest error in scenario 1 with The MAPE value is 5.6349444%, and if MAPE < 10% then prediction or forecasting very accurate. Scenario with lowest error value generated by scenario 2 with The MAPE value is 2.3001713%, and if MAPE < 10% then prediction or forecasting very accurate. So that predictions with scenario 2 more accurate from scenario 1 and if the more Lots number of inputs then predictions the more accurate. Keywords— Energy, Electricity, ANFIS, MAPE.]]>
Copyrights © 2024
