Wavelet Transform and its Application in Time Series Analysis of Various Foreign Exchange Rates
Dhananjay Kumar Gupta · 2015
Wavelet Transform based multi-resolution analysis has been used previously to study the time series from various areas of science and economics. Here we try to analyze various foreign exchange rates over the last 14 years with the help of wavelet transforms. We carefully study the variations to find out the various periodicities present in the data and correlation properties. For this purpose we use the Daubechies 4 and the Morlet wavelet to carry out the discrete and the continuous wavelet transforms respectively. We use the discrete wavelet transform to de-noise the series to some extent and remove some of the very high frequency components and get an approximation of the original dataset. We then use continuous wavelet transform on this approximation to analyze the behavior of this time series at different frequencies and extract the periodic nature from it.