کد مقاله کد نشریه سال انتشار مقاله انگلیسی نسخه تمام متن
438797 690329 2012 11 صفحه PDF دانلود رایگان
عنوان انگلیسی مقاله ISI
Generic subset ranking using binary classifiers
موضوعات مرتبط
مهندسی و علوم پایه مهندسی کامپیوتر نظریه محاسباتی و ریاضیات
پیش نمایش صفحه اول مقاله
Generic subset ranking using binary classifiers
چکیده انگلیسی

A widespread idea to attack the ranking problem is by reducing it into a set of binary preferences and applying well studied classification methods. In particular, we consider this reduction for generic subset ranking, which is based on minimization of position-sensitive loss functions. The basic question addressed in this paper relates to whether an accurate classifier would transfer directly into a good ranker. We propose a consistent reduction framework guaranteeing that the minimal regret of zero for subset ranking is achievable by learning binary preferences assigned with importance weights. This fact allows us to further develop a novel upper bound on the subset ranking regret in terms of binary regrets. We show that their ratio can be at most 2 times the maximal deviation of discounts between adjacent positions. We also present a refined version of this bound when only the quality over the top rank positions is of concern. These bounds provide theoretical support on the use of the resulting binary classifiers for solving the subset ranking problem.

ناشر
Database: Elsevier - ScienceDirect (ساینس دایرکت)
Journal: Theoretical Computer Science - Volume 456, 19 October 2012, Pages 89-99