PERANCANGAN APLIKASI ANDROID PERAMALAN KURS MENGGUNAKAN METODE AVERAGE-BASED FUZZY TIME SERIES MARKOV MODEL

Oleh :  Johari, Ro’fah Nur Rachmawati,Widodo Budiharto

The currency is one of the important roles in the future at today’s open economy, a predictive representation of the value of the currency rates can help to take a decision about financial problems. In this paper, predictive analysis of Currency Rates (USD-IDR) will be applied using a method AFTSM.. This method embraces 4 main concepts: Average Based concept, Fuzzy concept, Time Series concept, and Markov Model concept. Average Based is used to count the range from historical data. Fuzzy concept is used to classify variables. Time Series concept is used to observe the JCI closing price within a certain period of time. Whereas Markov Model concept is used in predicting process which exerts the transition probability matrix. To support the predicting analysis process, an Android based application has been designed to analyze a time series data set of currency rates, then to predict the future price of Currency Rates so that these processes can be done easily, fast, more efficient and conscientiously. The result of this research is a next working day Currency Rates prediction using Average Based Fuzzy Time Series Markov Model. The prediction price obtained is quite accurate and helps user to predict currency rates next day ahead for their economic activity even their financial problems.

Keywords: Average based, Fuzzy time series, Markov, Fuzzy, Currency, Android.