Volatility and correlation: Modeling and forecasting using Support Vector Machines

Several Realized Volatility and Correlation estimators have been introduced. The estimators which are defined based on high frequency data converge to the true estimators faster than their counterparts even under Market Microstructure Noise. Also a strategy for multivariate volatility estimation has been introduced. The strategy which is an incorporation of Support Vector Machine with Multiresolution Analysis based on wavelets affords higher performance of estimation than the single estimation.