Title: Deploying and evaluating multiattribute product recommendation in e-markets

Authors: Nikos Manouselis

Addresses: Informatics Laboratory, Division of Informatics, Mathematics and Statistics, Department of Science, Agricultural University of Athens, Greece

Abstract: Evaluation of proposed recommender systems before their final deployment, under experimental conditions close to the actual e-market ones, can prove of significant added value for system developers. The objective of this paper is to present the results of a pilot evaluation experiment for a product recommendation system of an electronic market (e-market). More specifically, it first presents the design of a product recommendation algorithm that is based on MultiAttribute Utility Theory (MAUT). Then it describes its implementation and evaluation for supporting consumers| wine selection in agricultural e-markets. This takes place through a small-scale experiment that engages a variety of evaluation measures and which involves both simulated data, as well as, data from a focus group of real wine consumers. This pilot evaluation may serve as a useful case study for e-market operators and developers who wish to evaluate multiattribute product recommendation systems before they deploy them online.

Keywords: agricultural e-markets; recommender systems; pilot study; evaluation; multiattribute utility theory; MAUT; wine selection; product recommendation; electronic markets.

DOI: 10.1504/IJMDM.2008.016041

International Journal of Management and Decision Making, 2008 Vol.9 No.1, pp.43 - 61

Published online: 03 Dec 2007 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article