Pendekatan Adaptive Neuro Fuzzy Sebagai Alternatif Bagi Bank Indonesia Dalam Menentukan Tingkat Inflasi Di Indonesia

Abstract

In uncertain economic like today, research and modeling the inflation rate is considered necessary to provide estimates and predictions of inflation rates in the future. Adaptive Neuro Fuzzy approach is a combination of Neural Network and Fuzzy Logic. This study aims to describe the movement of inflation (output variable ) so it can be estimated by observing four Indonesia's macroeconomic data, namely the exchange rate, money supply, interbank interest rates, and the output gap (input variable). Observation period started from the data in 20011 to 20113. After the learning process is complete, fuzzy systems generate 45 fuzzy rules that can define the input-output behavior. The results of this study indicate a fairly high degree of accuracy with an average error rate is 0.5315.