We propose an asymmetric GARCH in mean mixture model and provide a feasible way for option pricing within this general framework by deriving the appropriate risk neutral dynamics. We forecast out-of-sample prices of a large sample of options on the S&P 500 index from January 2006 through December 2011 and compute dollar losses and implied standard deviation losses. We compare our results to existing mixture models and other benchmarks like component models and jump models. Using the model confidence set test, the overall dollar root mean squared error of the best performing benchmark model is significantly larger than the best mixture model. Link to the article
ROMBOUTS, J. and STANTOFT, L. (2015). Option Pricing with Asymmetric Heteroskedastic Normal Mixture Models. International Journal of Forecasting, 31(3), pp. 635-650.