TY  - JOUR
TI  - A suggestion for constructing a Bayesian network model with simple correlation and an appropriate regression analysis: A real medical diagnosis application
AB  - The major task of medical science is to prevent or diagnose disease. Medicaldiagnosis is usually made by using some blood metrics and in addition, to be able toreach better results, one can benefit from different scientific methods. In this paper aBayesian network method is proposed. This method is a hybrid that uses simplecorrelation and according to dependent variable type either simple linear regression orlogistic regression for constructing a Bayesian topology. The Bayesian network is amethod for representing probabilistic relationships between variables associated with anoutcome of interest. To develop a Bayesian network, a structure must first beconstructed. To build the topology of the Bayesian network, some alternative methodcan be used. One is using domain experts who usually have a good grasp of theconditional dependencies in the domain to develop the structure of the Bayesiannetwork. Another is using structure learning algorithms, such as genetic algorithms, toconstruct the network topology from training data. In this paper a different constructionmethod is proposed by using correlation analysis and one of the simple linear regressionor logistic regression analyses. First, correlations of the examined variables are found.Then according to the significant correlation coefficients, the degree and direction of theinteractions between these variables are established by using either simple linearregression or logistic regression. Finally the Bayesian network model is constructed byusing this information. For evaluating our model, another model which does not haveany relation between the input variables is also constructed. And these two models arecompared by using an original thyroid data set. It is concluded that our proposed modelprovides a high degree of performance and good explanatory power and it may proveuseful for clinicians in the medical field.
AU  - ERPOLAT, Semra
PY  - 2012
JO  - Mathematical and Computational Applications
VL  - 17
IS  - 3
SN  - 1300-686X
SP  - 212
EP  - 222
DB  - TRDizin
UR  - http://search/yayin/detay/129976
ER  -