Thumbnail
Access Restriction
Open

Author Zacharia, Giorgos ♦ Evgeniou, Theodoros ♦ Boussios, Costas
Source CiteSeerX
Content type Text
File Format PDF
Subject Domain (in DDC) Computer science, information & general works ♦ Data processing & computer science
Abstract We develop a framework within which robust models of preferences are computationally efficiently estimated using quadratic optimization methods. Within this framework general highly non-linear models can be computationally efficiently estimated while at the same time avoiding overfitting problems that such models typically have. We compare these models with standard logistic regression and recently proposed polyhedral conjoint methods. Motivation: Traditional preference modeling methods such as conjoint analysis [2] have been used for many preference data analysis applications [12] but have been more naturally suited for controlled data gathering situations for example through questionnaires. However a lot of web-based information about choices is typically not gathered in such a controlled way and therefore is more noisy and often sparse. It is therefore important to develop a new generation of preference modeling methods that are robust to handle such data, especially for products with a very large number of attributes, and can be computed efficiently without sacrificing the complexity of the models. Previous Work: The market research community has traditionally approached utility estimation problems through function estimation. Conjoint analysis is one of the main methods for modeling preferences from data [2, 6]. For simplicity, we deal only with full-profile preference data- full product comparisons- for the
Educational Role Student ♦ Teacher
Age Range above 22 year
Educational Use Research
Education Level UG and PG ♦ Career/Technical Study