Implied Volatility Pricing with Selective Learning

Machine learning presents itself as an alternative data-driven approach to predict implied volatilities in Fintech. However, such an approach suffers from a relatively low prediction accuracy besides a model selection issue. In this study, we propose a novel selective learning approach to enhance machine learning implied volatility pricing. It boosts different machine learning models’ performance on different option data on behalf of moneyness, besides identifying optimal machine learning models in implied volatility prediction. In particular, selective learning can be an excellent way to enhance implied volatility pricing for the option datasets with more noise. In addition, we find out-of-the-money (OTM) and in-the-money (ITM) options fit machine learning prediction better than near-the-money (NTM) options. This pioneering work first provides a robust way to enhance implied volatility pricing via machine learning and will inspire similar studies in the future.