Self-Organising Maps Of Exchange Rates Through Hopfield Neural Network
by Krishnan Rengganathan
Abstract
Globalisation and liberalisation of economies had removed the barriers to trade and this has tied’ in enormous improvement in export and import activities as well as capital flows among counties. Sovereign funds, corporations and institutional investors are continuously seeking investment opportunities outside the country as domestic market appears to have reached a saturation and this retards growth and expansion. With globalisation, foreign exchange has become an important instrument of trade. Without having a profound understanding about the foreign exchange rates, the management of foreign currencies’ exposures gives rise to challenges and potential losses which are not to be accepted as fate. This risk which is inevitable o be properly managed and controlled for a healthy growth of foreign trade.
The study had applied secondary data obtained from the online portal of Pacific Exchange Rate Services. The analysis is performed using the exchange rates of eight major currencies, namely D EUR GBP, AUD, NZD, JPY, KRW and HKD for the period between 1 January 2009 and April 2012 (only with four months data). Later, it was established that HKD is pegged to the SD after they produced a perfect correlation and hence, HKD was dropped from further analysis.
The main objective of this research is to find the Self-Organising Maps (SOM) for each currency and to determine whether the exchange rate returns fall within the boundary rectangle for the years 2009, 2010, 2011 and 2012 for the selected currencies. The correlation coefficient computed on exchange rates for these selected currencies revealed equally positive and negative coefficient. Another notable finding was that the NZD was closely moving with the AUD. The same level of correlation coefficient was observed for these currencies even when the exchange ‘ales are converted to returns. When exchange rates are converted to returns, many correlation coefficient which were negative in exchange rates disappeared and became positive when they are converted to return.
Please request the Thesis from us via the button below.