Paper Details: Downloads: 2095
Serial Number: P1111818619
Title: Applying Support Vector Machine and AdaBoost Method to Ordinal Classification Problem
Authors: A.A. Onoja and G.O. Orwa and O.O. Ngesa3
Abstract: The paper discusses two approaches to classifying ordinal dataset problem in an imbalanced multi-class scenario which are daily eye-saw in human activities involving ordering. Due to model overfitting that often incur false or biased predictions which affect decision making because of imbalanced class instances and the importance of class sensitivity are crucial to researchers in identifying underlying patterns that are rare and not resolved easily using a single classifier, therefore there are need to validate the results from the single classifier using another classifier that is sensitive to class imbalances distributions. The paper, typically focused on the model performance recall and accuracy of the two approaches to handling classification problems in imbalanced multi-class ordinal dataset, comparing results from Support Vector Machine kernel functions (linear, normalized polynomial and Radial Basis function) and AdaBoost ensemble method. And ascertained the best approach that incorporate class sensitivity and account for model overfitting using practical examples of household food security status.
Keywords: Ordinal classification, Imbalanced Multiclass Dataset, SVM, AdaBoost, Model overfitting, Household food security status.
Journal/Conference: International Journal of Automatic Control and System Engineering
Volume: 18
Issue: 1
Submission Date: 4/28/2018 12:00:00 AM
Review Date: 6/18/2018 12:00:00 AM
Publishing Date: 7/6/2018 12:00:00 AM
Article Downloads: 2095
Download:

Facebook