Generalizations of the noisy-or model

Jiří Vomlel

Kybernetika (2015)

  • Volume: 51, Issue: 3, page 508-524
  • ISSN: 0023-5954

Abstract

top
In this paper, we generalize the noisy-or model. The generalizations are three-fold. First, we allow parents to be multivalued ordinal variables. Second, parents can have both positive and negative influences on their common child. Third, we describe how the suggested generalization can be extended to multivalued child variables. The major advantage of our generalizations is that they require only one parameter per parent. We suggest a model learning method and report results of experiments on the Reuters text classification data. The generalized noisy-or models achieve equal or better performance than the standard noisy-or. An important property of the noisy-or model and of its generalizations suggested in this paper is that it allows more efficient exact inference than logistic regression models do.

How to cite

top

Vomlel, Jiří. "Generalizations of the noisy-or model." Kybernetika 51.3 (2015): 508-524. <http://eudml.org/doc/271597>.

@article{Vomlel2015,
abstract = {In this paper, we generalize the noisy-or model. The generalizations are three-fold. First, we allow parents to be multivalued ordinal variables. Second, parents can have both positive and negative influences on their common child. Third, we describe how the suggested generalization can be extended to multivalued child variables. The major advantage of our generalizations is that they require only one parameter per parent. We suggest a model learning method and report results of experiments on the Reuters text classification data. The generalized noisy-or models achieve equal or better performance than the standard noisy-or. An important property of the noisy-or model and of its generalizations suggested in this paper is that it allows more efficient exact inference than logistic regression models do.},
author = {Vomlel, Jiří},
journal = {Kybernetika},
keywords = {Bayesian networks; noisy-or model; classification; generalized linear models; Bayesian networks; noisy-or model; classification; generalized linear models},
language = {eng},
number = {3},
pages = {508-524},
publisher = {Institute of Information Theory and Automation AS CR},
title = {Generalizations of the noisy-or model},
url = {http://eudml.org/doc/271597},
volume = {51},
year = {2015},
}

TY - JOUR
AU - Vomlel, Jiří
TI - Generalizations of the noisy-or model
JO - Kybernetika
PY - 2015
PB - Institute of Information Theory and Automation AS CR
VL - 51
IS - 3
SP - 508
EP - 524
AB - In this paper, we generalize the noisy-or model. The generalizations are three-fold. First, we allow parents to be multivalued ordinal variables. Second, parents can have both positive and negative influences on their common child. Third, we describe how the suggested generalization can be extended to multivalued child variables. The major advantage of our generalizations is that they require only one parameter per parent. We suggest a model learning method and report results of experiments on the Reuters text classification data. The generalized noisy-or models achieve equal or better performance than the standard noisy-or. An important property of the noisy-or model and of its generalizations suggested in this paper is that it allows more efficient exact inference than logistic regression models do.
LA - eng
KW - Bayesian networks; noisy-or model; classification; generalized linear models; Bayesian networks; noisy-or model; classification; generalized linear models
UR - http://eudml.org/doc/271597
ER -

References

top
  1. Almond, R. G., Mislevy, R. J., Steinberg, L., Yan, D., Williamson, D., 10.1007/978-1-4939-2125-6, Statistics for Social and Behavioral Sciences. Springer, New York 2015. MR3328529DOI10.1007/978-1-4939-2125-6
  2. Apt{é}, Ch., Damerau, F., Weiss, S. M., 10.1145/183422.183423, ACM Trans. Inform. Syst. 12 (1994), 3, 233-251. DOI10.1145/183422.183423
  3. Byrd, R. H., Lu, P., Nocedal, J., Zhu, C., 10.1137/0916069, SIAM J. Sci. Comput. 16 (1995), 1190-1208. Zbl0836.65080MR1346301DOI10.1137/0916069
  4. Díez, F. J., 10.1016/b978-1-4832-1451-1.50016-0, In: Proc. Ninth Conference on Uncertainty in Artificial Intelligence, Morgan Kaufmann 1993, pp. 99-105. DOI10.1016/b978-1-4832-1451-1.50016-0
  5. Díez, F. J., Druzdzel, M. J., Canonical Probabilistic Models for Knowledge Engineering., Technical Report CISIAD-06-01, UNED, Madrid 2006. 
  6. Díez, F. J., Galán, S. F., 10.1002/int.10080, Int. J. Intell. Syst. 18 (2003), 165-177. DOI10.1002/int.10080
  7. Heckerman, D., Breese, J., 10.1016/b978-1-55860-332-5.50041-9, In: Proc. Tenth Conference on Uncertainty in Artificial Intelligence, Seattle, Morgan Kaufmann 1994, pp. 286-292. DOI10.1016/b978-1-55860-332-5.50041-9
  8. Henrion, M., Practical issues in constructing a Bayes' Belief Network., In: Proc. Third Conference Annual Conference on Uncertainty in Artificial Intelligence, AUAI Press 1987, pp. 132-139. 
  9. Jensen, F. V., Nielsen, T. D., 10.1007/978-0-387-68282-2, Springer, 2007. MR2344166DOI10.1007/978-0-387-68282-2
  10. McCullagh, P., Regression models for ordinal data., J. Roy. Statist. Soc. Series B (Methodological) 42 (1980), 109-142. Zbl0483.62056MR0583347
  11. McCullagh, P., Nelder, J. A., 10.1007/978-1-4899-3242-6, Chapman and Hall, London 1989. Zbl0744.62098MR3223057DOI10.1007/978-1-4899-3242-6
  12. Miller, R. A., Fasarie, F. E., Myers, J. D., Quick medical reference (QMR) for diagnostic assistance., Medical Comput. 3 (1986), 34-48. 
  13. Neal, R. M., 10.1016/0004-3702(92)90065-6, Artif. Intell. 56 (1992), 1, 71-113. Zbl0761.68081MR1171969DOI10.1016/0004-3702(92)90065-6
  14. Pearl, J., Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference., Morgan Kaufman, San Mateo 1988. Zbl0746.68089MR0965765
  15. Team, R Development Core, R: A Language and Environment for Statistical Computing., R Foundation for Statistical Computing, Vienna 2008. 
  16. Rijmen, F., 10.1016/j.ijar.2008.01.001, Int. J. Approx. Reas. 48 (2008), 2, 659-666. Zbl1184.62039MR2419024DOI10.1016/j.ijar.2008.01.001
  17. Samejima, F., Estimation of Latent Ability Using a Response Pattern of Raded Scores (Psychometric Monograph No. 17)., Psychometric Society, Richmond 1969. 
  18. Saul, L. K., Jaakkola, T., Jordan, M. I., Mean field theory for sigmoid belief networks., J. Artif. Intell. Res. 4 (1996), 61-76. Zbl0900.68379
  19. Savický, P., Vomlel, J., Exploiting tensor rank-one decomposition in probabilistic inference., Kybernetika 43 (2007), 5, 747-764. Zbl1148.68539MR2376335
  20. Srinivas, S., 10.1016/b978-1-4832-1451-1.50030-5, In: Proc. Ninth Conference on Uncertainty in Artificial Intelligence, Morgan Kaufmann 1993, pp. 208-215. DOI10.1016/b978-1-4832-1451-1.50030-5
  21. Vomlel, J., 10.1002/int.20141, Int. J. Intell. Syst. 21 (2006), 381-398. Zbl1160.68584DOI10.1002/int.20141
  22. Vomlel, J., A generalization of the noisy-or model to multivalued parent variables., In: Proc. 16th Czech-Japan Seminar on Data Analysis and Decision Making under Uncertainty 2013, pp. 19-27. 
  23. Vomlel, J., Tichavský, P., On tensor rank of conditional probability tables in Bayesian networks., A preprint arXiv:1409.6287, 2014. MR3178417
  24. Zagorecki, A., Druzdzel, M. J., 10.1109/tsmca.2012.2189880, IEEE Trans. Systems, Man, and Cybernetics: Systems 43 (2013) 186-195. DOI10.1109/tsmca.2012.2189880

NotesEmbed ?

top

You must be logged in to post comments.

To embed these notes on your page include the following JavaScript code on your page where you want the notes to appear.

Only the controls for the widget will be shown in your chosen language. Notes will be shown in their authored language.

Tells the widget how many notes to show per page. You can cycle through additional notes using the next and previous controls.

    
                

                
Note: Best practice suggests putting the JavaScript code just before the closing </body> tag.