| Accurate prediction of market behavior has become a challenging task due to the inherently noisy and non-linear characteristics of the market.Using predictive values,you can price assets and make strategic decisions to get short-term or long-term bene-fits.There are various statistical predictors available on the market today,and different results can be obtained.This paper first reviews several options pricing methods including the Black-Scholes model and the basic principles of several types of machine learning algorithms.The four common machine learning algorithms will be followed:support vector machine mod-els;clustering and support vector machine hybrids.Model;input prediction and support vector machine hybrid model;reinforcement learning method,applied to the numeri-cal calculation of option pricing,and compared with the theoretical price obtained by Black-Scholes option pricing formula.This article mainly refers to the discussion of Deoda,A.[11]and Martin,K.[18],and further understanding,and then use the actual data to verify the method proposed in the paper.The comparison of the results of different methods shows that the hybrid model using clustering and S VR method is better than the simple SVM model.The method of using predictive input parameters performs poorly compared to other machine learning methods.Interestingly,the performance of most models has improved as we move from the ITM option series to the OTM option series.This can be attributed to the fact that the index tends to rise at a long-term price level.However,existing procedures are very time consuming because simulations must be generated for each option to be priced.In addition,traders need to predict prices more accurately to reduce risk in options trading. |
