The Modeling Research of Soybean Meal Option Implied Volatility Surface Based on the Linear Polynomial Model

Wen Yandong,Song Guijuan,Zhou Hongcheng,Yandong Wen,Guijuan Song,Hongcheng Zhou

DOI: https://doi.org/10.1109/icitbs.2019.00172

2019-01-01

Abstract:On the basis of studying the data characteristics of soybean meal option, in this article, we changed Dumas Linear Polynomial model, by enhancing the power of polynomial of parameter τ, to obtain several models. And based on the market data, we tested these models with different perspectives and the consequences are effective. From the result of the experiments, we can conclude four principles as followed. Firstly, with the complexity of model increased, the model can outperform on the side of data explanation. Secondly, it was going to present over-fitting if the model gets more complicated, which weakened the model's prediction ability. Thirdly, although the more complex models outperformed, the basic one with power of polynomial of parameter τ equaled one also conducted well in prediction. Fourthly, we developed that as the power of polynomial of parameter τequaled two, the algorithm presented better ranging from fitness to prediction, which is adapted to establish the model of soybean meal options with intraday quotation.