IJSRSET1844509 | Received : 01 May 2018 | Accepted : 08 May 2018 | May-June-2018 [(4) 8 : 49-53] © 2018 IJSRSET | Volume 4 | Issue 8 | Print ISSN: 2395-1990 | Online ISSN : 2394-4099 Themed Section : Engineering and Technology 49 Analysis of The Profitability of Islamic Banking Using Arimax Model and Regression with Arima Errors Model Sri Aryani* 1 , Muhammad Nur Aidi 2 , Utami Dyah Syafitri 3 * 1 Currently Pursuing Master Degree at Department of Statistics, Bogor Agricultural University, Bogor, West Java, Indonesia 2 Lecturer at Department of Statistics, Bogor Agricultural University, Bogor, West Java, Indonesia 3 Lecturer at Department of Statistics, Bogor Agricultural University, Bogor, West Java, Indonesia ABSTRACT Profitability of Islamic banking is a benchmark for the performance of Islamic bank. Indicators for measuring the profitability of Islamic banking is ROA. There are several predictor variables involved that affect the profitability of Islamic banking, namely interest rates, exchange rates and inflation. This research was aimed to predict the effect of predictor variables on Islamic bank profitability using two approaches namely ARIMAX and ARIMA errors. Both model approaches are conceptually the combination between the regression model and ARIMA model. The best model of the two approaches will be compared by considering the MAPE value. The best model is reflected by the smallest SBC value, significant parameter estimator, and the smallest MAPE value. As a result, the approach model of ARIMA (1, 1, 1) error with variable of exchange rate was found to be the best model with MAPE value of 32.013%, while the model of ARIMAX (0, 1, 2) with exchange rate variable obtained MAPE value of 375.78%. Thus, the best model to predict the effect of predictor factor on Islamic banking profitability was ARIMAX model (0,1,2) with exchange rate variable. Keywords : ARIMAX, ARIMA errors, Predictor, Profitability, ROA I. INTRODUCTION Building a model to predict the effect of correlation between independent variables and dependent variable can be done using time series model approach. Time series is a series of variables influenced by the time and at the same interval[1]. Time series can also be defined as the collection of data from an event in sequence based on the timeline of occurrences[2]. Banking profitability is one of studies conducted in the field of socio-economics that mainly applies the approach of linear regression method. This regression method approach is aimed to correlate the effect of dependent variable (Z) and independent variable (X), namely the effect of independent variable on profitability variable. Some previous researches that applied regression method approach include the study conducted by Abduh and Idress [3] which showed that inflation was found to have positive and significant impact on the profitability of Islamic Banks in Malaysia. Research of Swadayani and Kusumaningtias [4] showed that the variables of inflation, interest rates, exchange rates and money supply together resulted in a significant effect on the ROA of Islamic banking in Indonesia. A recent study conducted by Hidayati [5] indicated that inflation and exchange rates significantly affected the profitability of Islamic banks in Indonesia.