Studies in the volatility process of financial markets have focused more on volatility modeling aspects, using parametric assumptions.... Show moreStudies in the volatility process of financial markets have focused more on volatility modeling aspects, using parametric assumptions. Compared with the vast amount of research in parametric modeling, there is a lack of studies in nonparametric approaches in volatility forecasting and forecasting performance evaluation. This research intends to explore alternative approaches to forecasting volatility of financial returns, and evaluation of forecasting models. This research will employ grammatical evolution to propose a hybrid forecasting model that utilizes the benefits of parametric and genetic programming models. Furthermore, an alternative methodology to handle structural breaks in volatility is examined by utilizing an adaptive approach in dynamic environments. In an extensive empirical study, the proposed models will be compared with the other models widely used in the literature using statistical and economic tests. Specifically, as an alternative to the statistical performance evaluation measure, a nontraditional method derived from the idea of speculating on asset volatility will be employed to compare the performance as well as to assess economic usefulness of competing volatility models. The hybrid model provided superior forecasting performance than traditional techniques both on economic and statistical measures. PH.D in Management Science, December 2012 Show less
