A Collaborative Meta-Heuristic for Exchange Rate Prediction of Different Currencies

This paper focuses on forecasting of currency exchange rates of different countries in a distributed manner. In existing techniques, future exchange rate prognosis is performed by mining the financial data of a particular country. A proportion of data is deployed for training the proposed model whereas rest portion is used to validate the model. Both training and testing phase are carried out by feeding the features extracted from the past data of a single country only. Furthermore, the raw data to be processed is required to be located on a single machine. On the contrary, the proposed approach has employed data of different countries, trained and processed them in a distributed way. The prediction performance of the suggested technique is more efficient compared to that of conventional technique with added advantage of coping with distributed and different data from similar financial domain. Inherent mechanism of the prediction model also eliminates the data transfer time to assimilate dissimilar data at a single source location and execute the algorithm on it.