基于HMM-EGARCH的银行间同业拆放利率市场波动预测研究  

Shibor volatility forecast based on hidden Markov model EGARCH model

作　　者：林宇[1] 陈粘 陈宴祥[2] 

机构地区：[1]成都理工大学商学院,成都610059 [2]成都理工大学管理科学学院,成都610059

出　　处：《系统工程理论与实践》2016年第3期593-603,共11页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金面上项目(71171025);国家社会科学基金(12BGL024);四川省软科学研究计划项目(2014ZR0093);成都理工大学"金融与投资"优秀创新团队计划项目(KYTD201303)~~

摘　　要：针对中国金融市场呈现出的多波动状态的典型事实特征,以上海银行间同业拆放利率(Shibor)市场为研究对象,不仅引入隐马尔可夫模型(hidden Markov model,HMM)对其进行了波动状态预测,而且还引入HMM-EGARCH模型对其波动率进行了预测;最后使用成功率(success rate,SR)与平均绝对误差(mean absolute error,MAE)对预测波动状态进行检验,并且还采用标准统计误差函数对预测波动率进行检验.实证研究表明:高低两种波动状态就能够有效地刻画出Shibor市场的波动状态;HMM模型能够对Shibor市场进行较准确地波动状态预测,且更重要的是,HMM模型对高波动状态预测具有显著的优越性;HMM(2)-EGARCH模型能够有效地对Shibor市场进行波动率预测.In the paper, which draws upon Shanghai interbank offered rate （Shibor） market typical features of the fact that for the Chinese financial market volatility presents a multi-volatile state not only we forecasted the state of the Shibor volatility by introducing hidden Markov model （HMM）, but also applied HMM-EGARCH model to forecast the Shibor volatility. At last, we used success rate （SR） and mean absolute error （MAE） to test forecasting its volatility stste, also used standard statistical error function to test the predicting volatility. Our results indicate that only using high and low volatility state is able to effectively characterize the volatility state of Shibor. Eurthermore, HMM method can more accurately forecast the state of Shibor, and it is more important that HMM has more excellence for forecasting state of high volatility. HMM-EGARCH model can effectively predict the volatility of Shibor.

关 键 词：多波动状态 Shibor市场 HMM HMM-EGARCH 

分 类 号：F224[经济管理—国民经济]