On the Testability of Models with Missing Data

17th International Conference on Artificial Intelligence and Statistics

数学科学;计算机科学

Karthika Mohan ; Judea Pearl

PID  :  123853

来源: CEUR

【 摘 要 】

Graphical models that depict the process by which data are lost are helpful in recover ing information from missing data. We ad dress the question of whether any such model can be submitted to a statistical test given that the data available are corrupted by miss ingness. We present sufficient conditions for testability in missing data applications and note the impediments for testability when data are contaminated by missing entries. Our results strengthen the available tests for MCAR and MAR and further provide tests in the category of MNAR. Furthermore, we provide sufficient conditions to detect the ex istence of dependence between a variable and its missingness mechanism. We use our re sults to show that model sensitivity persists in almost all models typically categorized as