Option trading strategies efficiency analysis for MOEX assets with a machine learning approach.

Abstract

Option trading strategies efficiency analysis for MOEX assets with a machine learning approach.

Kudryavtsev O.Ye., Rodochenko V.V., Mamedzade Kh.M., Chivchyan A.A.

Incoming article date: 24.03.2017

An inportant and challenging task for a market analyst is to predict a value or a direction of a certain indicator based on the behaviour of a given financial instrument. In this paper we use random forest algorithms to predict a value and a direction of an implied volatility of RTS index options. We collect and prepare the historical data with an algorithm which takes into account both the actual trades history and the ordlog. We then examine and use the effect of volatility clustering which can be observed on such options. Finally, we construct some simple option trading strategies based on the predictions we made and measure their performance on a subset of our historical data. The results obtained shows that such approach can of use as one of the mechanisms for making decisions in practical trading, and needs further research.

Keywords: option, trading strategies, random forest, RTS index, spot market