Collaborative recommendation A robustness analysis推荐.pdfVIP

Collaborative Recommendation: A Robustness Analysis MICHAEL O’MAHONY, NEIL HURLEY, NICHOLAS KUSHMERICK and ´ ´ GUENOLE SILVESTRE University College Dublin Collaborative recommendation has emerged as an effective technique for personalised informa- tion access. However, there has been relatively little theoretical analysis of the conditions under which the technique is effective. To explore this issue, we analyse the robustness of collaborative recommendation: the ability to make recommendations despite (possibly intentional) noisy prod- uct ratings. There are two aspects to robustness: recommendation accuracy and stability. We formalise recommendation accuracy in machine learning terms and develop theoretically justified models of accuracy. In addition, we present a framework to examine recommendation stabil- ity in the context of a widely-used collaborative filtering algorithm. For each case, we evaluate our analysis using several real-world data-sets. Our investigation is both practically relevant for enterprises wondering whether collaborative recommendation leaves their marketing operations open to attack, and theoretically interesting for the light it sheds on a comprehensive theory of collaborative recommendation. Categories and Subject Descriptors: H.3.3 [Information Search and Retrieval]: Information Filtering General Terms: Experimentation, Performance, Reliability Additional Key Words and Phrases: Collaborative recommendation, machine learning, robustness 1. INTRODUCTION Collaborative recommendation has emerged as an effective personalisation tech- nique for a diverse array of electronic commerce and information access scenarios, e.g. Goldberg et al. [1992]; Shardanand and Maes [1994]. Such systems keep track of their customers’ preferences, and use these data to offer new suggestions. Many