In the last years, there has been a growing interest in the analysis of three- way (matrix-variate) data via mixture models. Quite often real data are affected by outliers,andthisalsooccursinthecontextofthree-waydata.Acommonsolutionfor managing this type of data consists in fitting mixtures of heavy-tailed distributions. Unfortunately, the three-way literature is still limited, with the only matrix variate 푡 mixtures recently proposed to cope with this issue. Therefore, in this work we firstly introduceanewheavy-tailedmatrix-variatedistributionandthenweuseitwithinthe mixture model setting. The resulting mixture model is finally fitted to a real dataset for illustrative purposes.
|Titolo:||A new mixture model for three-way data|
|Data di pubblicazione:||2020|
|Appare nelle tipologie:||4.1 Contributo in Atti di convegno|