A ranking system for comparing models' performance combining multiple statistical criteria and scenarios: The case of reference evapotranspiration models. (April 2019)

• Record Type: Journal Article
• Authors: Aschonitis, V.G.; Lekakis, E.; Tziachris, P.; Doulgeris, C.; Papadopoulos, F.; Papadopoulos, A.; Papamichail, D.
• Is Part Of: Environmental modelling & software. Volume 114(2019)
• Page Start: 98
• Page End: 111
• Publication Date: 2019-04
• Subjects: Statistical criteria -- Relative distance of models -- Criteria selection -- Cluster analysis -- Reference crop evapotranspiration models; Environmental monitoring -- Computer programs -- Periodicals; Ecology -- Computer simulation -- Periodicals; Digital computer simulation -- Periodicals; Computer software -- Periodicals; Environmental Monitoring -- Periodicals; Computer Simulation -- Periodicals; Environnement -- Surveillance -- Logiciels -- Périodiques; Écologie -- Simulation, Méthodes de -- Périodiques; Simulation par ordinateur -- Périodiques; Logiciels -- Périodiques; Computer software; Digital computer simulation; Ecology -- Computer simulation; Environmental monitoring -- Computer programs; Periodicals; Electronic journals; 363.70015118
• DOI: 10.1016/j.envsoft.2019.01.005
• Languages: English
• ISSNs: 1364-8152

Abstract

Abstract: The aim of this study is to propose a new system for ranking models' performance. The system is based on the standardized Ranking Performance Index ( sRPI ), which combines the results of any number and type of criteria from different simulation scenarios. The sRPI ranges between 0 and 1 (1 for the best and 0 for the worst performing model) based on the relative distance of models' accuracy. The sRPI is also combined with an additional index called RCI (Ranking Criteria Index) that identifies which criteria are the most and the least representative for a specific dataset/scenario. The presentation of sRPI methodology is performed comparing ten reference evapotranspiration models using a large list of statistical criteria for both daily and monthly steps. The sRPI methodology could significantly reduce the complexity to evaluate models' performance based on multiple scenarios and can identify the most descriptive criteria reducing their overuse in modelling studies. Highlights: A new method for ranking models' performance is proposed. The method combines any number of statistical criteria and simulation scenarios. The method ranks the models based on their relative distance. The method also identifies the most descriptive criteria for a specific dataset.