基于AR模型的股指结构化特征研究  

Research on Structural Characteristics in Stock Index Based on AR Model

作　　者：许长龙[1] 蔡世民[1] 周佩玲[1] 

机构地区：[1]中国科学技术大学电子科学与技术系,合肥230026

出　　处：《计算机工程》2009年第10期240-242,共3页Computer Engineering

基　　金：国家自然科学基金资助项目(70571075)

摘　　要：应用AR模型分析股价趋势变化的二元符号序列,得到该二元符号序列具有线性模型的特征,通过有限阶的条件概率分析该二元符号序列,得到该股指序列具有明显的结构化特征,从不同的角度揭示金融市场的不完全有效性。通过对比新兴和成熟股指的结构化特征,发现新兴的金融市场确实存在明显的非理性投资行为。The changing trends of binary sequences are analysed with AR model（the autoregressive model）, to obtain the linear model characteristics. And through a limited order of conditional probability analysis of the binary sequences, the stock index has a remarkable structural feature. This paper is from another perspective to reveal that the financial markets are not completely effective. By comparing the structural characteristics of the stock index of the developing and developed markets, it finds that the non-rational investment behavior is in the immature emergent financial markets.

关 键 词：结构化特征 AR模型 条件概率 二元符号序列 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]