摘要
结构突变是统计学、经济学、信号处理和生物信息学等学科领域中的研究热点之一.Z.Harchaoui等提出了基于LASSO的结构突变点检测方法,是近几年结构突变问题的最新研究方法.为了在国内推行该方法,系统介绍了国外基于LASSO方法的几种变点模型中的变点检测问题,其核心是把变点检测问题转化成模型选择问题来解决,并阐述了相应的算法.最后探讨该方法在不同学科领域的应用和前景展望.
Structural break problem is one of the hot topics in statistics,economics,signal processing and bioinformatics and other fields. Harchaoui and Levy-Leduc(2008) initiatively proposed structural break point detection method based on LASSO method,which is a new method in dealing with structural break problems. The paper systematically introduces the change point detection problem based on LASSO method in several change point models.The core is transforming the change point detection problem into model selection problem. The corresponding algorithms are introduced. Finally,the applications and perspective of this new method in some fields are put forward.
出处
《江西师范大学学报(自然科学版)》
CAS
北大核心
2015年第2期189-193,共5页
Journal of Jiangxi Normal University(Natural Science Edition)
基金
全国统计科学研究重点课题(2011LZ035)
山东省自然科学基金(ZR2014AL006)
上海财经大学研究生创新基金(CXJJ-2014-445)资助项目