Bagging and Boosting with Equivalent Space Consumption

Agarwal, Rishabh and Mishra, Satwik and Arjun, C V (2017) Bagging and Boosting with Equivalent Space Consumption. In: International Conference on Contemporary issues in Science, Engineering & Management, 28/05/2017, Hotel Excellency, Bhubaneswar.

[img] PDF
960.pdf - Published Version
Restricted to Registered users only

Download (2MB) | Request a copy


There are various bagging techniques of which Random Forests and Extra Trees are among the most popular. Similarly, AdaBoost and Gradient Boosting are the most common boosting techniques. There is a lot of discussion as to which machine is better and why and which one should be used with different experts having different opinions and the choice based mostly on the problem and personal choice. But the question is how to effectively compare them since bagging and boosting machines, although being ensemble methods, construct their component estimators differently. This paper aims to compare their respective accuracy and time requirement when all the four considered machines use the same number of trees, with all of these trees having the same size by applying them on a randomly generated uniformly distributed data set to remove any chance of a particular machine providing better results on some standard data set.

Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: Bagging, Boosting, Equivalent Space, Outliers, Trees
Subjects: Engineering > MIT Manipal > Information and Communication Technology
Depositing User: MIT Library
Date Deposited: 11 Dec 2017 10:09
Last Modified: 11 Dec 2017 10:09

Actions (login required)

View Item View Item