A market-based approach to recommender systems

Abstract

Recommender systems have been widely advocated as a way of coping with the problem of information overload for knowledge workers. Given this, multiple recommendation methods have been developed. However, it has been shown that no one technique is best for all users in all situations. Thus we believe that effective recommender systems should incorporate a wide variety of such techniques and that some form of overarching framework should be put in place to coordinate the various recommendations so that only the best of them (from whatever source) are presented to the user. To this end, we show that a marketplace, in which the various recommendation methods compete to offer their recommendations to the user, can be used in this role. Specifically, this article presents the principled design of such a marketplace (including the auction protocol, the reward mechanism, and the bidding strategies of the individual recommendation agents) and evaluates the market's capability to effectively coordinate multiple methods. Through analysis and simulation, we show that our market is capable of shortlisting recommendations in decreasing order of user perceived quality and of correlating the individual agent's internal quality rating to the user's perceived quality.

User evaluation of a market-based recommender system

“…For each such user-selected recommendation, the suggesting agent is given a reward. In defining the ComputeReward function, our aim is to ensure that it is both Pareto efficient and social welfare maximizing (as motivated in [13]). Since the global objective is to shortlist the most valuable recommendations in decreasing order of relevance, as perceived by the user, we decided to reward the user-selected recommendations based on this feedback.…”

“…where δ and α are two system coefficients (δ > 0 and α > 1) and P M+1 is the highest not shortlisted bid price (the detailed justification for this particular choice is given in [13]). The values of δ and α will depend upon the specifics of the application, but they need to be set at suitable values to ensure R h > P h so that the rewarded agents can make profits.…”

“…Such bids can result in one of the following outcomes occurring: the bid is not shortlisted, it is shortlisted but not rewarded, or it is rewarded. With respect to a given inq level, a bidder's strategy depends on the last outcome in the following way (again see [13] for a justification for these choices):…”

“…To this end, we have previously specified a system that recommends relevant online documents (represented as urls) in which this coordination is achieved via a marketplace (see Figure 1) in which recommender agents compete with one another to have their suggestions placed before the user. See [13] for a detailed justification of the choice of a market-based approach for this problem and for a detailed comparison with the state of the art in recommender and multi-agent systems.…”

Computational Service Economies: Design and Applications

“…In the former case, we have made advances in the areas of auctions (Dash et al, 2007Rogers et al, 2007a;Vetsikas et al, 2007;Gerding et al, 2008), coalition formation (Dang and Jennings, 2006;Fatima et al, 2009;Rahwan and Jennings, 2007;Chalkiadakis et al, 2008;, automated negotiation (Fatima et al, 2006;Karunatillake et al, 2009;Ramchurn et al, 2007;Fatima et al, 2004), trust and reputation , flexible reasoning strategies for workflows (Stein et al, 2009a) and decentralized coordination (Rogers et al, 2007b, Farinelli et al, 2008. In the latter case, we have built applications using these techniques in areas such as: virtual organizations (Norman et al, 2004), emergency response (Chapman et al, 2009), sensor networks (Padhy et al, 2006;Kho et al, 2009;, mobile sensors (Stranders et al, 2009), computational grids (Stein et al, 2009b) and personalized recommendations (Wei et al, 2005;Payne et al, 2006).…”