基于神经网络方法的期权定价研究

Study on the Option Pricing Based on ANN Method

人工神经网络是一种非线性非参数模型,它不对市场和变量做任何的主观假设。所以它具有许多存在假设条件的模型所无法具备的优点。通过选择输入和输出变量,设置BP神经网络的结构,选择预测期限就可以训练该神经网络模型,训练成功后即可对我国的权证进行预测。并利用四个误差指标来评价模型定价结果的优劣,实证研究发现神经网络方法在我国权证市场的定价效果要好于B-S定价方法。

Artificial Neural Network is a non-linear non-parametric model and has no Assumption. It has high-speed computing and learning characteristics, shows its superiority in modeling complex systems especially in the areas such as prediction and evaluation. We apply the ANN model to pricing warrants after fix the input and output, the structure and the longs it predicate. It found that the ANN model has a better prediction than the B-S model according to the four error indicators.