Recent Development of Distance-Based Models for Ranking Data

• Activity Number: 281
• Type: Topic Contributed
• Date/Time: Tuesday, August 5, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Philip Yu*+ and Paul H. Lee and Fang Qi
• Companies: University of Hong Kong and Hong Kong Polytechnic University and University of Hong Kong
• Keywords: Ranking data ; Kendall distance ; Spearman distance ; Decision tree ; Nearest neighbors methods

Abstract

Analysis of ranking data arises from different fields of study, such as psychology, economics, and politics. Many models for ranking data have been developed in the recent decades. Among them, distance-based ranking models postulate that the probability of observing a ranking of items depends on the distance between the ranking and a modal ranking. However, under a distance-based model, the closer a ranking to the modal ranking, the higher the ranking probability is. In other words, the distribution of rankings is single-peaked at the modal ranking, which is not reasonable for heterogeneous data. In this talk, we introduce some recent development of distance-based models for ranking data including distance-based tree models and local distance-based models. Empirical studies will be conducted to assess their prediction accuracy.