Role of Attribute Selection in Classification Algorithms
S. Dinakaran, P. Ranjit Jeba Thangaiah · 2013
Feature selection is the technique of removing irrelevant features and to reduce dimensionality of the feature. This paper proposed attribute selection of information gain attribute evaluator and ranker search method to selected attribute and each selected attribute is ranked based on the filter and wrapper method. Tree based J48 classifier is used with different test options namely 10 fold cross validation, use training set, supplied test set, and percentage split default of 66% are compared with all options and to generate best accuracy results; Labor dataset is implementing to test the above methods.