Related Experiment Videos
The influence on partial order ranking from input parameter uncertainty. Definition of a robustness parameter
P B Sørensen1, B B Mogensen, L Carlsen
1Department of Environmental Chemistry, The National Research Institute in Denmark, Roskilde. pbs@dmu.dk
Chemosphere
|May 20, 2000
Summary
This study analyzes how uncertainty in environmental data affects partial order ranking. Results show ranking uncertainty increases significantly when robustness parameter E drops below 4-5 comparisons per element.
Area of Science:
- Environmental Science
- Data Analysis
- Ranking Methodologies
Background:
- Partial order ranking is valuable for environmental data but is sensitive to data uncertainty.
- Environmental datasets frequently contain significant levels of uncertainty, complicating analysis.
Purpose of the Study:
- To analyze the general influence of data uncertainty on partial order ranking in environmental contexts.
- To develop a method for quantifying and understanding ranking uncertainty due to data variability.
Main Methods:
- A Monte Carlo simulation approach was employed using randomly generated datasets.
- Partial order rankings were converted to a one-dimensional scale, accounting for varying certainty.
- A general robustness parameter (E), representing expected comparisons per element, was defined and correlated with uncertainty.
Main Results:
- Data uncertainty significantly impacts partial order ranking results.
- Ranking uncertainty escalates sharply as the robustness parameter (E) falls below 4-5 comparisons per element.
- Beyond E=5, ranking uncertainty stabilizes and becomes largely independent of E's specific value.
Conclusions:
- The robustness parameter (E) provides a quantifiable measure of ranking reliability under data uncertainty.
- Environmental data uncertainty can be managed by ensuring an adequate number of comparisons per element (E > 5).
- The findings offer a practical approach to interpreting the reliability of partial order rankings in environmental studies.