Abstract:
Classification is a form of data analysis that can be used to extract models describing import data class or to predict future data trends. This paper contains two important aspects of pattern recognition that classification problem and evaluate the performance are studied on real-world datasets. This system is to study the Naive Bayesian classifier and to classify the class label of two datasets. In this system, classifier is built on the training dataset and test the unknown dataset on the testing dataset. And then, calculate the accuracy of classifier by using the hold-out method. Before the classifier is built, missing value is filled by using mean value and feature value is normalized by using min-max normalization such as preprocessing step. The experiment is performed on two medical datasets from University of California, Irvine (UCI) machine learning database and General Hospital in Mandalay.
