Resumen
The purpose of this paper is to construct the model of the financial crisis in Indonesia through exchange market pressure index approach by using Multivariate Adaptive Regression Spline. This research used secondary data from the Central Bank of Indonesia from 2005 to 2014, consisting of 120 observations. The dependent variables is exchange market pressure index, and the independent variables consist of 11 macro economics variable. This research used the MARS 2.0 software, to build the model. The results shows 53.9% accuracy model of MARS and it obtains the smallest value of GCV that is 1.84, and the international interest rate of US Prime Rate is the most influential variable towards the exchange market pressure index. The results also provide additional knowledge regarding the indicators that can lead to the financial crisis based on the model established by the MARS approach. The implication is that the variable of international interest rate of US Prime Rate through the MARS approach can be an early warning system against the crisis that probably will happen, especially in Indonesia.