JSM 2012 Home

JSM 2012 Online Program

The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.

Online Program Home

Abstract Details

Activity Number: 617
Type: Contributed
Date/Time: Thursday, August 2, 2012 : 8:30 AM to 10:20 AM
Sponsor: Business and Economic Statistics Section
Abstract - #304685
Title: Learning from Only Positive and Unlabeled Data Using Generalized Stochastic Frontier Model
Author(s): Liuxia Wang*+
Companies: Sentrana Inc.
Address: 1725 I St. NW, Suite 900, Washington, DC, 20006, United States
Keywords: learning from only positive and unlabeled data ; stochastic frontier model ; Bayesian
Abstract:

One common problem in marketing science is to predict whether a certain customer needs to purchase certain items. However, instead of all items a customer has purchased from any stores, we only have incomplete information about the purchased items from one particular store most of the time. Predicting customers' need given such incomplete data is closely related to the unlabeled data problem in machine learning where the training data consists of a set of positive examples and a set of unlabeled examples, some of which are positive and some of are negative. In the paper, customers need given positive and unlabeled data is modeled using a generalized stochastic frontier model. Stochastic frontier model is an econometric model which adds a latent inefficiency term to the traditional linear regression model to explicitly capture the inefficiency in unobserved purchases from other stores. We generalize the stochastic frontier model to handle non-Gaussian errors and thus can be used to capture both the predictable purchase pattern and unobserved purchase patterns. We also illustrate this method using simulated examples, which demonstrates that it outperforms some existing methods.


The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2012 program




2012 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.