Outliers Data Mining in Normal-Inverse Gaussian Model

Author

Wang, Li-li ; Hou, Xiang-yang ; Xiong, Yan-ye

Author_Institution

Sch. of Command Autom., PLA Univ. of Sci. & Technol., Nanjing, China

Volume

1

fYear

2010

fDate

4-6 June 2010

Firstpage

231

Lastpage

234

Abstract

The normal-inverse model arises as a normal variance-mean mixture with an inverse Gaussian mixing model. The resulting model, it is very complicated to obtain the influence measures based on the tradition method. In the present paper, several diagnostic measures for outlier data mining are obtained based on the conditional expectation of the complete-data log-likelihood function based on the EM algorithm. An example for which we apply the diagnosis methods is given as illustration.

Keywords

Gaussian processes; data mining; expectation-maximisation algorithm; EM algorithm; complete-data log-likelihood function; diagnostic measures; normal variance-mean mixture; normal-inverse Gaussian model; outliers data mining; Algorithm design and analysis; Automation; Data mining; Density measurement; Educational institutions; Electronic mail; Inverse problems; Probability density function; Programmable logic arrays; Signal processing algorithms; EM algorithm; generalized Cook distance; local influence analysis; normal inverse Gaussian model;

fLanguage

English

Publisher

ieee

Conference_Titel

Information and Computing (ICIC), 2010 Third International Conference on

Conference_Location

Wuxi, Jiang Su

Print_ISBN

978-1-4244-7081-5

Electronic_ISBN

978-1-4244-7082-2

Type

conf

DOI

10.1109/ICIC.2010.65

Filename

5514191