On the Testability of Models with Missing Data

Karthika Mohan, Judea Pearl
Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics, PMLR 33:643-650, 2014.

Abstract

Graphical models that depict the process by which data are lost are helpful in recovering information from missing data. We address the question of whether any such model can be submitted to a statistical test given that the data available are corrupted by missingness. 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 existence of dependence between a variable and its missingness mechanism. We use our results to show that model sensitivity persists in almost all models typically categorized as MNAR.

Cite this Paper


BibTeX
@InProceedings{pmlr-v33-mohan14, title = {{On the Testability of Models with Missing Data}}, author = {Mohan, Karthika and Pearl, Judea}, booktitle = {Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics}, pages = {643--650}, year = {2014}, editor = {Kaski, Samuel and Corander, Jukka}, volume = {33}, series = {Proceedings of Machine Learning Research}, address = {Reykjavik, Iceland}, month = {22--25 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v33/mohan14.pdf}, url = {https://proceedings.mlr.press/v33/mohan14.html}, abstract = {Graphical models that depict the process by which data are lost are helpful in recovering information from missing data. We address the question of whether any such model can be submitted to a statistical test given that the data available are corrupted by missingness. 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 existence of dependence between a variable and its missingness mechanism. We use our results to show that model sensitivity persists in almost all models typically categorized as MNAR.} }
Endnote
%0 Conference Paper %T On the Testability of Models with Missing Data %A Karthika Mohan %A Judea Pearl %B Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2014 %E Samuel Kaski %E Jukka Corander %F pmlr-v33-mohan14 %I PMLR %P 643--650 %U https://proceedings.mlr.press/v33/mohan14.html %V 33 %X Graphical models that depict the process by which data are lost are helpful in recovering information from missing data. We address the question of whether any such model can be submitted to a statistical test given that the data available are corrupted by missingness. 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 existence of dependence between a variable and its missingness mechanism. We use our results to show that model sensitivity persists in almost all models typically categorized as MNAR.
RIS
TY - CPAPER TI - On the Testability of Models with Missing Data AU - Karthika Mohan AU - Judea Pearl BT - Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics DA - 2014/04/02 ED - Samuel Kaski ED - Jukka Corander ID - pmlr-v33-mohan14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 33 SP - 643 EP - 650 L1 - http://proceedings.mlr.press/v33/mohan14.pdf UR - https://proceedings.mlr.press/v33/mohan14.html AB - Graphical models that depict the process by which data are lost are helpful in recovering information from missing data. We address the question of whether any such model can be submitted to a statistical test given that the data available are corrupted by missingness. 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 existence of dependence between a variable and its missingness mechanism. We use our results to show that model sensitivity persists in almost all models typically categorized as MNAR. ER -
APA
Mohan, K. & Pearl, J.. (2014). On the Testability of Models with Missing Data. Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 33:643-650 Available from https://proceedings.mlr.press/v33/mohan14.html.

Related Material