Priya.S, Antony Selvadoss Thanamani
Predictive classification as a wide range of application in data mining. Most real data set have missing values which affects the accuracy of classifiers. This paper will investigate predictive performance of missing data using two classifier techniques naive Bayes classifier and Knn classifier. Among the two classifiers naive bayesian is least sensitive and provides a good accuracy to handle missing data but K nearest neighbour is the most sensitive to missing data. NB is one of the classifiers that handle missing data very well, it just excludes the attribute with missing data when computing posterior probability (i.e. probability of class given at a data point).
Classifiers, Naive Bayes Classifier And Knn Classifier, Predictive.