Header menu link for other important links
Application of bagging and boosting for all the classification algorithms
A. Singh,
Published in
Volume: 8
Issue: 3
Pages: 17669 - 17680
The Classification algorithms in Data mining are the backbone for classifying and predicting the nominal datasets. These techniques are used to find out patterns between data attributes and results in probalistic prediction of the label attribute. Ensembling is a technique used to improve the results of the algorithms. The paper uses bagging and adaboosting to enhance the classification results but it is not always necessary to have better outcomes. Three classification algorithms named Naive Bayes,Random Forests and Decision Trees are used to classify two different nominal datasets. Both the datasets are selected from the repository sites according to the number of intances. CM1 dataset has less number of instances than Adult-Census dataset. This paper compares and analizes the three algorithms on two datasets and results out that the Naïve Bayes is most suitable for small datasets and Decision Tree is suitable for large datasets. Further ensembling techniques are applied and analysis are done for each algorithm on each dataset using RapidMiner tool while equating precision,recall and accuracy. Boosting failed for CM1 dataset but showed its worth for Adult-Census dataset. Whereas bagging has mostly improved results on all analysis and has better results than adaboost. © 2016,International Journal of Pharmacy and Technology. All rights reserved.
About the journal
JournalInternational Journal of Pharmacy and Technology