Abstract
Nonlinear data smoothers provide a practical method of finding smooth traces for data confounded with possibly long-tailed or occasionally “spikey” noise. While they are natural tools for analyzing time-series data, they can be applied to any data set for which a sequencing order can be established. Their resistance to the effects of unsupported extreme observations and their ability to respond rapidly to well-supported patterns make them valuable as tools for finding patterns not constrained to specific parametric form and as versatile data-cleaning algorithms. This paper defines some robust nonlinear smoothers that have performed well in Monte-Carlo trials and makes brief recommendations based upon that study.
Keywords: Monte-Carlo trials, time-series data, data noise reduction, nonlinear filters, exploratory data analysis
Full text
PDF
