TIME SERIES ANALYSIS WITH BACKPROPAGATION NEURAL NETWORK AND ADAPTIVE NEURO FUZZY INFERENCE SYSTEM
Zarina Ulfa
Department of Statistics, Faculty of Mathematics and Natural Sciences
Islamic University of Indonesia

ABSTRACT

Inflation is one of indicator to reflect economic stability in Indonesia. High inflation makes market prices rise and low inflation makes the value of money decline, therefore inflation stabilitiy must be maintened. One way that can be done is to do a time series analysis. However, past events must be considered, for example is hyperinflation (uncontrolled inflation). Hyperinflation makes data to be non-liniear so that time series analysis is needed that can understand non-linear data patterns for forecasting. Time series analysis used is Backpropagation Neural Network (BPNN) and Adaptive Neuro Fuzzy Inference System (ANFIS). From the analysis, it was found that the ANFIS method is more suitable to used to predict the inflation with an MAPE error is 8,78% for training and 9,64% for testing.
Keywords: Inflation, BPNN, ANFIS, MAPE