预测传播结构与人工金融市场的易变性

Spread Structure of Predication and Anti-Durative of the Artificial Financial Market

为了探索价格预测信息传播动力结构与市场价格波动特征之间的关系,建立了一个基于元胞自动机的人工金融市场模型。在模型中,引入了动态有向图来表征异质投资者之间信息传播的邻居关系网络;投资者个体在邻居影响下进行价格预测与投资决策,从而形成市场价格。通过对仿真交易价格时间序列的R/S分析,发现邻居关系网络结构对于投资者预测的同质化有重要作用;投资者预测同质化倾向强的资本市场时,价格时间序列的易变性随之增强,市场价格趋于不稳定。

In order to study the relationship between price fluctuation and transmission dynamic structure of the predictions, this article presents an artificial stock market model based on cellular automaton. In this model, dynamic directed graph was utilized to describe neighborhood networks of predication transmission in heterogeneous investors. Individual investors predict market price and make investment decision, which determine market price, under the influences of their neighbors. By R/S analysis of the time series of the prices produced by simulated exchange, it is found that neighborhood network structure is of great important to the homogenization of the investors＇ predictions. In market with powerful homogenization of the investors＇ predictions, anti-durative of the time series of the prices increases and market price tends to be unstable.