Modification of Tukey’s smoothing techniques for extreme data

Two Tukey’s techniques which are resistant line for linear trend and resistant smoothing for non linear trend have been reviewed in this research. The new resistant line for method of dividing the batches using range dealing with ties and non ties is recommended. The determination of sample size...

Full description

Saved in:
Bibliographic Details
Main Author: Husain, Qasim Nasir
Format: Thesis
Language:English
Published: 2017
Online Access:http://psasir.upm.edu.my/id/eprint/67711/1/FS%202017%2084%20%20IR.pdf
http://psasir.upm.edu.my/id/eprint/67711/
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Two Tukey’s techniques which are resistant line for linear trend and resistant smoothing for non linear trend have been reviewed in this research. The new resistant line for method of dividing the batches using range dealing with ties and non ties is recommended. The determination of sample size in each batch is also being introduced. In resistant smoothing, mathematical terms have been initiated and incorporated in this technique, the part which has been neglected by introducer of exploratory data analysis. New symmetric mean, right mean, left mean, right median, and left median have been proposed, leading to more simple process of smoothing technique. Later, the proposed methods have been used for the suggested compound smoothing techniques and hannings with simpler, faster and better smooth. Additionally, in order to evaluate the efficiency of variant proposed techniques together with the smoothing index and extra balance test, simulation data with big data size have successfully been applied.