Service Selection Based on Similarity Measurement for Conditional Qualitative Preference

Similarity measurement is essential in many preference-based personalized applications such as collaborative recommendation and service selection. Up to date, current researches have mainly focused on the measurements for quantitative preference rather than qualitative preference, although the latter has attracted much attention recently. Only a very few methods to measure user similarity are proposed. This paper aims to fill in this gap by proposing an intuitive similarity measurement for conditional qualitative preference which is represented by CP-nets. Experimental results based on two expanded real-life datasets demonstrate that our similarity measurement is not only able to correctly reflect user´s preference changes, but also effective to identify similar users.