Statistical performance comparison between the FSV and the FS-NMI index
Document typeConference report
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Rights accessRestricted access - publisher's policy
This paper presents a performance comparison between two validation methods developed specifically for the Computational Electromagnetics purposes: the Feature Selective Validation (FSV) and the Feature Selective Normalized Mutual Information (FSNMI) index. To achieve this goal, a statistical analysis of 40 different cases of study (pairs of data sets) is carried out covering a wide range of real-life applications, such as, frequency domain, noisy and transient data, among others. The results provided an insight of the relationships between each method, showing that more effort is required to achieve generally coherent validation results between the FSV and the FSNMI.
CitationAzpúrua, M. [et al.]. Statistical performance comparison between the FSV and the FS-NMI index. A: IEEE International Symposium on Electromagnetic Compatibility. "EMC 2014 Proceedings : IEEE International Symposium on Electromagnetic Compatibility : 3-8 August, Raleigh, North Carolina". Raleigh, North Carolina: Institute of Electrical and Electronics Engineers (IEEE), 2014, p. 322-327.
|STATISTICAL PER ... V AND THE FS-NMI INDEX.pdf||1.114Mb||Restricted access|