A0260
Title: On the matrix-variate normal distribution for interval-censored and missing data
Authors: Salvatore Daniele Tomarchio - University of Catania (Italy)
Salvatore Ingrassia - University of Catania (Italy)
Antonio Punzo - University of Catania (Italy)
Victor Hugo Lachos Davila - University of Connecticut (United States) [presenting]
Abstract: Matrix-variate distributions are powerful tools for modeling three-way datasets that often arise in longitudinal and multidimensional spatiotemporal studies. However, observations in these datasets can be missing or subject to some detection limits because of the restriction of the experimental apparatus. A novel matrix-variate normal distribution for interval-censored and/or missing data is proposed. An analytical yet efficient EM-type algorithm is developed to conduct maximum likelihood estimation of the parameters, having closed-form expressions that rely on truncated moments. Results obtained from the analysis of both simulated data and real case studies concerning water quality monitoring are reported to demonstrate the effectiveness of the proposed method.
