非参数可加ACD模型及其应用研究

An Research on the Nonparametric Additive ACD Model and the Application in the Stock Market of China

非参数可加ACD模型对条件期望的函数形式与随机误差项的分布形式要求都没有参数ACD模型强,因此不会像参数ACD模型那样因模型形式设定错误而得出错误结论。非参数可加ACD模型估计出来的各个可加部分图形的形状对于正确设定参数ACD模型具有一定的指导作用。

We have proposed the nonparametric additive ACD model and carried on the demonstration analysis based on the simulation sample and the China Stock market＇s adjusted price duration sample to the model in the paper. The function form of the conditional expectation and the distribution form of the stochastic error term of the nonparametric additive ACD model have fewer limits than the parametric ACD model. So the model will not make wrong conclusion because of the wrong model form hypothesis just like parametric ACD model. This point may obtain the confirmation in our demonstration analysis. Nonparametric additive ACD model can estimate the graph of each additive component, which has some help to specify the form of parametric ACD model.