Improving Classification Accuracy Using Ensemble Learning Technique (Using Different Decision Trees)

Using ensemble methods is one of the general strategies to improve the accuracy of classifier and predictor. Bagging is one of the suitable ensemble learning methods. Ensemble learning is a simple, useful and effective metaclassification methodology that combines the predictions from multiple base classifiers (or learners). In this paper we show a comparative study of different classifiers (Decision trees) when the ensemble learning technique called bagging is used. We perform classification on various datasets firstly by using a single classifier and then by bagging method, using the same base classifier. It is observed that when we use a single classifier rather than an ensemble, the classification error further increases. As different training data subsets are randomly drawn-with-replacement from the entire training dataset, so usually the new training set contains some duplicates and some omissions as compared to the original training set. Each training data subset is used to train a different classifier of the same type.