Bayesian Gaussian process factor analysis with copula for count data

Gregor Pirs, Erik Strumbelj

Abstract: Highlights•New method for smooth latent trajectory estimation for multivariate count data.•Proposed method accounts for residual covariance, not explained by latent factors.•Estimated latent trajectories are robust to noise.•Interpretable simultaneous forecasting of several disease counts.•Robust estimation of National Basketball Association teams latent strengths.