您的位置:首页 > 论文页面
股票交易市场中Cox比例风险模型的应用
发表时间:2015-07-15 浏览量:2705 下载量:1551
全部作者: | 杨淮,赵学靖 |
作者单位: | 兰州大学数学与统计学院 |
摘 要: | 将Cox比例风险模型应用于股票交易市场,旨在获得影响股票生存时间的重要影响因素。以上证100指数所对应的基本成分股为研究对象,观测时间为2014年1月至2014年4月。首先,自定义股票的生存时间,得出每支股票的生存时间和生存状态;其次,分析影响每支成分股交易价格的13个财务指标两两之间的相关系数,结果显示,某些协变量之间的相关性较强,而另一些协变量之间的相关性较弱,因此对协变量进行变量选择非常必要;然后,利用坐标下降算法对带有惩罚函数的最小绝对值压缩与选择(least absolute shringkage and select eprator, Lasso)型偏似然函数进行计算,再用交叉检验方法选出合适的惩罚系数,得到影响股票生存时间的重要协变量;最后,利用似然比检验法进行参数检验,取检验水平0.05,检验通过,说明利用Lasso方法选择的影响股票生存时间的重要协变量是合理的。 |
关 键 词: | 数理统计学;Cox比例风险模型;变量选择;交叉检验;上证100指数 |
Title: | Application of Cox proportional hazards model in stock market |
Author: | YANG Huai, ZHAO Xuejing |
Organization: | School of Mathematics and Statistics, Lanzhou University |
Abstract: | This paper devotes to the application of the Cox model in stock market in order to choose the covariates which have important influence to the stock. We studied the Shanghai Stock Exchange (SSE 100) index between January 2014 to April 2014. The survival time and state are obtained by defining the time of failure. The coefficients between 13 finance indexes are explored to indicate the interdependency, which shows the necessity of variables selection. Using the minimization of least absolute shringkage and select eprator (Lasso) type partial likelihood with penalty function by coordinating descent algorithm and cross-validation method for turning parameter, the important covariates are obtained. The significance of the parameters is tested by likelihood ratio test under the predetermined significance level of 0.05. The numerical analysis demonstrate the effectiveness of the Lasso method in variables selection in stock market. |
Key words: | mathematical statistics; Cox proportional hazards model; variable selection; cross-validation; Shanghai Stock Exchange (SSE 100) index |
发表期数: | 2015年7月第13期 |
引用格式: | 杨淮,赵学靖. 股票交易市场中Cox比例风险模型的应用[J]. 中国科技论文在线精品论文,2015,8(13):1344-1350. |

请您登录
暂无评论