Personalization in Mobile and Pervasive Computing Workshop Summary
D. Cheng,A. Kobsa,Kinshuk,K. Partridge,Zhiwen Yu
2009-01-01
Abstract:Providing the right information and services to the right person at the right time has been a hot research topic for many years, and great progress has been made in search quality, targeted advertisement, and content recommendation. But with the mobile handheld device becoming a dominant platform for accessing the Web, and smart objects springing up in many places, mobility and pervasiveness have compounded the difficulties in personalization, and made situation awareness even more important. This workshop brings together researchers and practitioners to discuss current and emerging research results and topics in personalization in mobile and pervasive computing. The nine papers of this workshop focus on three topics: 1) the unique issues, difficulties, and potential solutions for personalization in mobile and pervasive computing, 2) the roles of domain knowledge and ontology for personalization, and efficient and effective ways of employing them, and 3) the possible solutions to personalization in an environment filled with smart objects. The papers will be published at The paper " Situation-aware User Interest Mining on Mobile Handheld Devices " by D. Cheng et al presents a novel unsupervised clustering algorithm for extracting situation-aware patterns from usage logs. The algorithm automatically estimates near-optimal number of clusters and cluster centroids. It models situation by taking advantage of sensors. 5-fold cross validations using real-world data show that the algorithm delivers higher accuracy than existing algorithms with much lower complexity. The paper also describes the research problems in situation-aware personalization that must be addressed before users can benefit from the learning algorithms and speculates on possible approaches to the solutions. The paper " Actionable User Intentions for Real-Time Mobile Assistant Applications " by T. Panagos et al describes a class of mobile assistant applications that provide users with suggestions and reminders based on derived user intentions in addition to time, location, and resource availability. The user intentions are derived from the entries of calendars and to-do lists by first identifying the ontology associated with an event then selecting the appropriate taxonomy within the ontology.