A study on the modified components of Asian Currency Unit: an application of the Artificial Neural Network

This paper identifies the modified components of the central value of the Asian Currency Unit (ACU). The purpose of this research was to utilize the Back-Propagation Neural Network (BPNN) to implement the modified components analysis and compared the results with that of the Radial Basis Function Neural Network (RBFNN). Eight models were constituted to evaluate the macroeconomic variables’ extent of influence on the central value of the ACU. The empirical evidence supports the need of the sample countries to construct the central rate of the ACU. By considering foreign direct investment (FDI), external debt, and bank claims from the private sector, the central value of the ACU could be expressed more effectively.