WITHDRAWN: Adaboost Classification of Ranking Data

• Author(s): Fang Qi and L.H. Yu
• Keywords: Ranking data ; Adaboost

Abstract

Ranking data has applications in different fields of studies, like politics, marketing. The classification of ranking data is more difficult than traditional classification problem because the preference of ranking will produce more classes (It will produce N factorial classes if N classes need to be ranked). Many models of ranking data have been developed over the years, like the distance-based model, decision tree model. Multi-class Adaboost has been a very successful technique for the classification problem with more than two-class. This algorithm combines weak classifier and will improve the performance greatly. Therefore, we will combine the traditional classifier with the boosting method in order to get a better classifier. The algorithms will be modified to suit for the ranking data, and we will compare the results of the traditional method and new method.