Abstract
In this work we introduce a methodology based on Genetic Algorithms for the automatic induction of Bayesian Networks from a file containing cases and variables related to the problem. The methodology is applied to the problem of predicting survival of people after one, three and five years of being diagnosed as having malignant skin melanoma. The accuracy of the obtained model, measured in terms of the percentage of well-classified subjects, is compared to that obtained by the so-called Naive-Bayes. In both cases, the estimation of the model accuracy is obtained from the 10-fold cross-validation method.
We thank Gregory F. Cooper for providing his simulation of the ALARM Network. We also thank the referees for their work and comments. This work was supported by the DiputaciĆ³n Foral de Gipuzkoa, under grant OF 92/1996, by the grant UPV 140.226-EA186/96 from the University of the Basque Country, and by the grant PI 95/52 from the Gobierno Vasco ā Departamento de EducaciĆ³n, Universidades e InvestigaciĆ³n.
Preview
Unable to display preview. Download preview PDF.
References
Andersen, S.K., Olesen, K.G., Jensen, F.V. and Jensen, F.: āHUGIN ā a shell for building Bayesian belief universes for expert systemsā. In Eleventh International Joint Conference on Artificial Intelligence (1989) 1128ā1133
Beinlinch, I.A., Suermondt, H.J., R.M. Chavez R. M. and Cooper G.F.: āThe ALARM monitoring system: A case study with two probabilistic inference techniques for belief networksā. In Proceedings of the Second European Conference on Artificial Intelligence in Medicine (1989) 247ā256
Bouckaert, R.R.: āOptimizing causal orderings for generating DAGs from dataā. In Uncertainty in Artificial Intelligence. Proceedings of the Eighth Conference (1992) 9ā16
Bouckaert, R.R.: āProperties of Bayesian belief networks learning algorithmsā. In Uncertainty in Artificial Intelligence. Tenth Annual Conference, (1994) 102ā109
Cooper, G.F., and Herskovits, E.A.: āA Bayesian method for the induction of probabilistic networks from dataā. Machine Learning 9 (1992) 309ā347
Goldberg, D.E.: āGenetic Algorithms in Search, Optimization and Machine Learningā. Addison-Wesley, Reading, MA (1989)
Heckerman, D., Geiger, D. and Chickering, D.M.: āLearning Bayesian networks: The combination of knowledge and statistical dataā. Technical Report MSR-TR-94-09, Microsoft (1994)
Izarzugaza, M.I.: āInforme del registro de CĆ”ncer de Euskadi 1990ā. Osasunkaria (1994) 8ā11
Jensen, F. V.: āIntroduction to Bayesian networksā. University College of London (1996)
LarraƱaga, P., Poza, M., Yurramendi, Y., Murga, R., and Kuijpers, C.: āStructure Learning of Bayesian Networks by Genetic Algorithms: A Performance Analysis of Control Parametersā. IEEE Transactions on Pattern Analysis and Machine Intelligence. 18 (1996) 912ā926
LarraƱaga, P., Murga, R., Poza, M., and Kuijpers, C.: āStructure Learning of Bayesian Networks by Hybrid Genetic Algorithmsā. In Learning from Data: AI and Statistics V, Lecture Notes in Statistics 112. D. Fisher, H.-J. Lenz (eds.), New York, NY: Spriger-Verlag, (1996) 165ā174.
LarraƱaga, P., Kuijpers, C., Murga, R., and Yurramendi, Y.: āLearning Bayesian Network Structures by searching for the best ordering with genetic algorithmsā. IEEE Transactions on System, Man and Cybernetics 26 (1996) 487ā493
Lauritzen, S.L.: āGraphical Modelsā. Oxford University Press (1996)
Lauritzen, S.L., and Spiegelhalter, D.J.: āLocal computations with probabilities on graphical structures and their application on expert systemsā. Journal Royal of Statistical Society B 50 (1988) 157ā224
Pearl, J.: āProbabilistic Reasoning in Intelligent Systems: Networks of Plausible Inferenceā. Morgan Kaufmann, San Mateo (1988)
Provan, G.M., and Singh, M.: āLearning Bayesian Networks Using Feature Selectionā. In Learning from Data: AI and Statistics V, Lecture Notes in Statistics 112. D. Fisher, H.-J. Lenz (eds.), New York, NY: Spriger-Verlag (1996) 291ā300
Robinson, R. W.: āCounting unlabeled acyclic digraphsā. In C. H. C. Little (ed.) Lectures Notes in Mathematics 622: Combinatorial Mathematics V, Springer-Verlag, New York, (1977) 28ā43.
Stone, M.: āCross-validation choice and assessment of statistical proceduresā. Journal Royal of Statistical Society 36 (1974) 111ā147
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
Ā© 1997 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
LarraƱaga, P., Sierra, B., Gallego, M.J., Michelena, M.J., Picaza, J.M. (1997). Learning Bayesian Networks by Genetic Algorithms: A case study in the prediction of survival in malignant skin melanoma. In: Keravnou, E., Garbay, C., Baud, R., Wyatt, J. (eds) Artificial Intelligence in Medicine. AIME 1997. Lecture Notes in Computer Science, vol 1211. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0029459
Download citation
DOI: https://doi.org/10.1007/BFb0029459
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-62709-8
Online ISBN: 978-3-540-68448-0
eBook Packages: Springer Book Archive