Wavelet-based prediction of crude oil prices

There are two kinds of transactions in the crude oil markets; one is based on immediate delivery while the other one on future delivery. The spot market is dependent on the rst kind of transactions and the future market is associated to the second one. Market condition ( e.g. market risk, irrational trading, etc. ) along with other factors ( e.g. credit risk, insurance risk, seasonal factors and etc. ) is often the main cause of uncertainty in the crude oil markets. Therefore, the future markets ( leading markets ) are built up to provide a cover structure for these uncertainties. Also crude oil future contracts, determine denitive prices in future deadlines to buy or sell according to specic criteria of delivery and payment. On the other hand, future prices re ects the markets expectations about future conditions. Consequently, large dierences between futures and spot prices is often used to describe the overall market conditions. Wavelets are used as a legitimate alternative alternative for irregular situations such as data or signals with scaled features, or containing discontinuities and sharp edges and so on (see [1-2]). In this study, we are going to use the wavelets as a suitable tool to investigate its performance in the crude oil futures markets (see [3]). We intend to provide forecasts over dierent forecasting horizons by introducing a prediction procedure and predicting future prices based on the wavelets by utilizing a series of data from the crude oil market and at last putting the results in comparison with the crude oil future markets data