Robust approaches for monitoring logistic regression profiles under outliers
International Journal of Quality & Reliability Management
ISSN: 0265-671X
Article publication date: 3 April 2017
Abstract
Purpose
The purpose of this paper is to develop some robust approaches to estimate the logistic regression profile parameters in order to decrease the effects of outliers on the performance of T2 control chart. In addition, the performance of the non-robust and the proposed robust control charts is evaluated in Phase II.
Design/methodology/approach
In this paper some, robust approaches including weighted maximum likelihood estimation, redescending M-estimator and a combination of these two approaches (WRM) are used to decrease the effects of outliers on estimating the logistic regression parameters as well as the performance of the T2 control chart.
Findings
The results of the simulation studies in both Phases I and II show the better performance of the proposed robust control charts rather than the non-robust control chart for estimating the logistic regression profile parameters and monitoring the logistic regression profiles.
Practical implications
In many practical applications, there are outliers in processes which may affect the estimation of parameters in Phase I and as a result of deteriorate the statistical performance of control charts in Phase II. The methods developed in this paper are effective for decreasing the effect of outliers in both Phases I and II.
Originality/value
This paper considers monitoring the logistic regression profile in Phase I under the presence of outliers. Also, three robust approaches are developed to decrease the effects of outliers on the parameter estimation and monitoring the logistic regression profiles in both Phases I and II.
Keywords
Acknowledgements
The authors are grateful to the anonymous referees for precious comments which led to improvement in the paper.
Citation
Hakimi, A., Amiri, A. and Kamranrad, R. (2017), "Robust approaches for monitoring logistic regression profiles under outliers", International Journal of Quality & Reliability Management, Vol. 34 No. 4, pp. 494-507. https://doi.org/10.1108/IJQRM-04-2015-0053
Publisher
:Emerald Publishing Limited
Copyright © 2017, Emerald Publishing Limited