ABSTRACT
In Statistical modeling, especially of financial series, the familiar and most known models are gaussian and log-normal ones. The gaussian models are based on the price increment hypothesis, while the log-normal ones are based on the log-price increment hypothesis. Methods based on multi-scale wavelet transformation provide powerful analysis tools that decompose time series data into coefficients related to time and a specific frequency band. Wavelets are considered capable of isolating fundamental low-frequency dynamics from non-stationary time series, and are robust to the presence of noise. Wavelet analysis allows the representation of time series into species relative to the time and frequency information known as time-frequency decomposition. The transformation into discrete wavelets uses circular filtering.
