Development of a visual rating scale to aid validations

摘要

Validation of electromagnetic modelling methods, particularly against experimental results often involves structurally complex data sets. The paper aims to establish a benchmarking technique, to aid the "visual" comparisons made between experimental and simulated data. The technique simply aids the human comparisons, and so does not remove the opinion and experience of the analyst making the comparison. The technique also provides the analyst with an overall quantitative measure of similarity. The detail of the development and improvements made to the modified scale is reported, and the effectiveness of the improved technique with regards to its ability to reduce the variance between analysts is emphasised. The paper concludes with highlighting the need for a "visual benchmark" and also emphasises the usefulness and effectiveness of the technique.