Signal Processing for Financial Markets

Background

In the last decades, there has been an explosive growth in the research area relating signal processing and financial trading (Gradojevic & Gencay, 2011; Bekiros, 2011), especially in the fields of business and banking researches and applications (Taskaya & Ahmad, 2003; Nuti et al., 2011; Zhang & Kedmey, 2011). Signal processing applications, which hold promising potential, remain relatively unexplored within finance but its applications for financial data would add a new perspective and could further enrich financial research. One example of signal processing used in finance is shown in (Benedetto et al., 2007, 2009), where the authors propose a business model founded on Bayes decision test for video-call billing based on the effective end-to-end quality of service (QoS). More recently, the IEEE Journal of Selected Topics in Signal Processing has promoted a Special Issue in “Signal Processing Methods in Finance and Electronic Trading” in August 2012 (Akansu et al., 2012), as well as the IEEE Signal Processing Magazine that has published a Special Issue in “Signal Processing for Financial Applications” in September 2011 (Pollak et al., 2011). However, empirical evidence shows that using traditional statistical and econometric models might lead to very poor prediction performances (Weigend & Gershenfeld, 1994). To this aim, the authors in (Abramson & Finizza, 1995) utilized a probabilistic model for predicting oil prices, while a semi-parametric statistical method is suggested in (Morana, 2001) for short-term oil price forecasting based on the GARCH properties of crude oil price. Similarly, a semi-parametric approach for oil price forecasting is introduced in (Barone-Adesi et al., 1998), while in (Yu et al., 2008) an empirical mode decomposition (EMD) based neural network ensemble learning paradigm is proposed for world crude oil spot price forecasting.