Comparative study on multivariate outlier detection methods in sesame (Sesamum indicum L.)
Abstract
Outlier detection in multivariate dataset is not quite trivial when compared to univariate. The tediousness in multivariate outlier is due to presence of swamping and masking effect which portrays an ideal sample point as outlier instead of true one. To overcome all this problems, robust techniques can be applied instead of classical outlier detection methods because the latter fails to find out the correct outlier. This paper enumerates various techniques like Mahalanobis, Cook’s, Leverage points, DFFITS, minimum volume ellipsoid (MVE) and minimum covariance determinant (MCD) for detection of outliers or anomalies in multivariate space and best will be identified. Researchers can use that technique to identify outliers before going for analysis, as this will assist in significant results.
It is certified that:
- The corresponding author is fully responsible for any disputes arising due to the publication of his/her manuscript.
- The article has been seen by all the authors who are satisfied with its form and content.
- The sequence of names of authors in the by-line is as per their relative contribution to this experiment, giving due credit to all scientists who made notable contribution to it.
- All the authors fully understand that inclusion of any other co-authors or exclusion of any co-authors is not possible once the article has been submitted to the journal.
- The corresponding author takes full responsibility for this article.
- The address of the organization where the research was conducted is given.
- The article is exclusive for this journal, and the results reported here have not been sent (and will not be sent during its consideration by this journal) for publication in any other journal.
- Authors agree to abide by the objective comments of referees and do agree to modify the article into a short note as per the recommendation, for publication in the Electronic Journal of Plant Breeding.
- If published in Electronic Journal of Plant Breeding, the copyright of this article would vest with the Indian Society of Plant Breeders, who will have the right to enter into any agreement with any organization in India or abroad engaged in reprography, photocopying, storage and dissemination of information contained in it, and neither we nor our legal heirs will have any claims on royalty.
