Detection of outliers in a multivariate linear regression model using Gibbs observation.
Abstract<br /> This paper deals with finding outliers in a multivariate linear regression model after assuming a model of normal - Wishart distribution. This method is based on the estimation of probability of an outlier for each observation by mixed Bernoulli model with shifting location outl...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Book |
| Published: |
College of Education for Pure Sciences,
2013-03-01T00:00:00Z.
|
| Subjects: | |
| Online Access: | Connect to this object online. |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract<br /> This paper deals with finding outliers in a multivariate linear regression model after assuming a model of normal - Wishart distribution. This method is based on the estimation of probability of an outlier for each observation by mixed Bernoulli model with shifting location outlier. We show how to obtain the posterior distribution in the mixed model by Gibbs sampler algorithm. Also the determination of the number of outliers is done by criterion of marginal likelihood distribution. The theoretical results of this research are applied to real data of multivariate linear regression. The results obtained are so encouraging in determining the outliers in these data. |
|---|---|
| Item Description: | 1812-125X 2664-2530 10.33899/edusj.2013.89420 |