Improving Accuracy of Classification System using Ensemble Technique

Nikhil Jadhav; Vaibhav Joshi; Bhushan Bhavsar; Vishwajeet Bokan; Amrut A Patil

Volume : VIII, Issue : VI, June - 2018

Improving Accuracy of Classification System using Ensemble Technique

Nikhil Jadhav, Vaibhav Joshi, Bhushan Bhavsar, Vishwajeet Bokan, Amrut A Patil

Abstract :

Ensemble techniques creates base classifiers and different individual output, after combining individual output one combined output is prepared usually by majority voting. To get better results by classifiers we use the technique called as Ensemble classification or Ensemble learning. Using Ensemble techniques is one of the general strategies to improve the accuracy and prediction. Ensemble Learning is simple, useful effective method that will combine prediction from multiple output of individual classifiers. Weather forecasting is a vital application in meteorology and has been one of the most scientifically and technologically challenging problems around the world in the last century. Ensemble classifiers in general targets to improve the accuracy and prediction multiple classifiers output to a single combined output with the help of majority voting. Boosting and bagging are the two techniques in the Ensemble Learning Boosting is the technique of combining group of individual output to strong output wherein Bagging is learning algorithm where multiple prediction are generated and subsequently aggregated by averaging.Email users are increasing day by day and they are facing the problem due to spam Emails. To filter the spam email there are using the Data mining techniques for classifying the Email data. They have used the ensemble of C5.0, SVM and ANN which are outperforming with high accuracy of 94.26%.