Abstract:
In this paper, we present a data-driven ensemble approach for option price prediction whose derivation is based on the no-arbitrage theory of option pricing. Using the theoretical treatment, we derive a common representation space for achieving domain adaptation. Through a specific scaling, suitable for financial time series data, we obtain a feature representation that is indistinguishable for samples coming from different domains. This provides an advantage over conventional models when predicting atypical out-of-sample test data. The success of an implementation of this idea is shown using some real market data. The root mean squared error in prediction turns out to be less than one-third of that for the benchmark model. We further report several experimental results for critically examining the predictive performance of the derived pricing models.