Classifier Ensembling: Dataset Learning Using Bagging and Boosting

Abstract

Data mining is used to tackle the huge amount data which are preserved in the data warehouse and database, to chase desired information. Many data mining have been projected such as association rules, decision trees, neural networks, etc. from many years it takes its position in center. The efficiency of the real time or artificial is being enhanced using certain data mining techniques like bagging and boosting and AdaBoost by which the accuracy of the different classifiers is also efficient. Boosting and bagging are two widely used ensemble methods for classification. As the task is to increase the efficiency of the classifier, as per the past works it is always better to have combination of classifiers than to go for random guessing. From the available boosting techniques, the AdaBoost algorithm is one of the best techniques especially in the case when carrying out the branching type of tasks. In the following work, a classification technique is explained which considers the ensemble of classifiers for bagging and for boosting separately, where decision tree is used as the classifier for bagging and for boosting artificial neural network (ANN) is being used as the classifier. For the ensembling, the classifier MDT is being used. Sections are divided as Introduction, Literature Review, Problem Statement, Issues and Challenges, Research Methodology, and Analysis of the Work.