Fast Estimation of the Modal Position for Unimodally Distributed Data
Title:
Fast Estimation of the Modal Position for Unimodally Distributed Data
Author:
Milatovic, Milan Badiru, Adedeji B.
Appeared in:
Intelligent data analysis
Paging:
Volume 2 (2013) nr. 1 pages 19-30
Year:
2013-06-14
Contents:
This study proposes a new approach for fast estimation of the position of the modal concentration, and consequently, its value, drawn from a unimodally distributed and sorted set of data. In terms of only five specific percentile values, the proposed procedure estimates the position of the modal concentration of unimodal curves with a relatively high accuracy. These five percentile values (in particular, 5th, 16th, 50th, 84th, and 95th) can be easily accessed from any given sorted distribution, regardless of the number of records in the database. The whole approach is based on the formula for Inclusive Graphic Skewness. Besides being one of the essential statistical parameters, the mode and the information about its location may be useful in other domains, such as searching non-uniform distributions, qualitative description of data, etc.
Journal description