Modeling Quantiles

View Sample PDF

Author(s): Claudia Perlich (IBM T.J. Watson Research, USA), Saharon Rosset (IBM T.J. Watson Research, USA)and Bianca Zadrozny (Universidade Federal Fluminense, Brazil)
Copyright: 2009
Pages: 6
Source title: Encyclopedia of Data Warehousing and Mining, Second Edition
Source Author(s)/Editor(s): John Wang (Montclair State University, USA)
DOI: 10.4018/978-1-60566-010-3.ch205

Purchase

View Modeling Quantiles on the publisher's website for pricing and purchasing information.

Abstract

One standard Data Mining setting is defines by a set of n observations on a variable of interest Y and a set of p explanatory variables, or features, x = (x1,...,xp), with the objective of finding a ‘dependence’ of Y on x. Such dependencies can either be of direct interest by themselves or used in the future to predict a Y given an observed x. This typically leads to a model for a conditional central tendency of Y|x, usually the mean E(Y|x). For example, under appropriate model assumptions, Data Mining based on a least squares loss function (like linear least squares or most regression tree approaches), is as a maximum likelihood approach to estimating the conditional mean. This chapter considers situations when the value of interest is not the conditional mean of a continuous variable, but rather a different property of the conditional distribution P(Y|x), in particular a specific quantile of this distribution. Consider for instance the 0.9th quantile of P(Y|x), which is the function c(x) such that P(Y

The IRMA Community

Research IRM

Modeling Quantiles

Purchase

Abstract

Related Content

IRMA Sponsors