(A) Study on rank-order based classification under the assumption of positive association

Abstract

Decision support system(DDS) is a computerized for helping make decisions between alternatives. Classification is a form of decision making under a certain loss structure due to time-efficiency. We are interested in a classification problem where the class levels are in accordance with the rank order of the probability values of a random variable with various background distribution. We consider a model-bases DSS where all the variables involved are binary, the probability that a certain variable is equal to 1(1 for success and 0 for failure)is categorized, and classifications are made for the variable in term of category levels. And we introduce a similary measure between Bayesian network(BN) models and describe how a BN model can be constructed which is similar to a given BN model. Then under the condition that all the variables are positively associated with each other, a method of obtaining the agreement levels between two models is described. Conclude by saying that we recommend to use this type of model to get robust classification when the Bayesian network model satisfies the positive association condition among the variables involved in the model. We assigned rank- order the conditional probabilities according to various Beta distributions to each variable which is provided from a Bayesian network model under PA condition. The conditional probability is the probability that a subject has a certain attribute given an outcome of some other variables and the classification is based the rank-order. When the Bayesian network model satisfies the positive association condition among the variables involved in the model, We can get robust classification with assuming same distribution for each variable, even which is uniform.