Ensembling Base Classifiers to Improve Predictive Accuracy

The algorithm of ensembling base classifiers can improve predictive accuracy, and achieve a better generalization. However, the ensemble classification methods in literature have been used in more rule-based algorithms of classifier. This paper presents a novel algorithm: CVCEEP (Classification by Voting Classifiers based on Essential Emerging Patterns). By learning the method of Bagging, multiple base-classifiers were generated on different bootstrap samples and combined as a powerful classifier by voting. Experimental results show that CVCEEP achieve a better predictive accuracy and can be match to the classic classification algorithms that we have known.