Wavelet Based Exchange Rate Forecasting with Improved Instance Based Learning

In this paper we present a novel wavelet based exchange rate forecast model integrating wavelet filters for denoising and Improved Instance Based Learning(IIBL) approach. The proposed model implements a novel technique that extends the nearest neighbor algorithm to include the concept of pattern matching so as to identify similar instances thus implementing a nonparametric regression approach.The work demonstrates the feasibility of integrating with suitable non-redundant orthogonal wavelet filters at the preprocessing stage to achieve accurate forecasting. The multi-scaling property of the wavelet transform enhances the prediction with high accuracy for volatile time series. The impact of using Discrete Wavelet Transform (DWT) has been systematically illustrated in the preprocessing stage on the accuracy of forecasting. The analysis of simulations demonstrate that the proposed wavelet based IIBL model results in accurate predictions and encouraging results for exchange rate series when compared with the conventional neural network, wavelet and wavelet denoising methods.