Accurate Parameter Elimination Approach in Ensemble Learning on NSL-KDD Dataset in Comparison with Candidate Elimination Algorithm

Aim: The primary importance of this research is to generate the comparison of a precise parameter elimination method in Ensemble understanding on NSL-KDD dataset in comparison with Candidate Elimination Algorithm. Materials and Methods: Accuracy is analyzed for feature elimination. Classification of Novel feature elimination is experimented by ensemble model of sample size (Number of samples=34) and Candidate Elimination model of (Number of samples=34) which produced with the G-power value 80%. Results: Ensemble model has achievement of 82.74% accuracy which has higher than Candidate Elimination Algorithm proves its accuracy of 74.38%. The significance results of accuracy is 0.095 p less than 0.05 indicates the performance of proposed work has insignificance. Conclusion: Ensemble model is better in finding accuracy of 82.74% when refers to comparison of Candidate Elimination Algorithm of accuracy 74.38% along with the hypothesis area to determine the set of features.