Open Conference Systems, CLADAG2023

Font Size: 
Jason Pillay, Andriette Bekker, Johan Ferreira, Mohammad Arashi

Last modified: 2023-07-06


Modelling noisy data in a network context remains an unavoidable obstacle. More so, network environments are most comprehensively described using random matrix theory, and thus necessitates the probabilistic characterisation of these networks (and accompanying noisy data) using matrix variate models. Denoising network data using a Bayes approach is not that common in surveyed literature, thus we briefly introduce a new matrix-variate T model in a prior sense for the noise process following the Gaussian graphical network, for the cases when the assumption of normality is violated in the model and cases when normality is no longer sufficient to explain variation in the data. Exploration of this matrix-variate T distribution's performance applied to such a network setting within a Bayesian context is considered. Calculation and approximation of the resulting posterior is of interest to assess the considered model’s betweenness and closeness, and is illustrated using real stock price data.