Personalized Smart Suggestions for Context-aware Human-activity Support by Ubiquitous Computing Networks

Accession number;04A0185876
Title;Personalized Smart Suggestions for Context-aware Human-activity Support by Ubiquitous Computing Networks
Author; SAKAI H (Ntt Information Sharing Platform Lab., Musashino-shi, Jpn) SUEDA Y (Ntt Information Sharing Platform Lab., Musashino-shi, Jpn) MURAKAMI K (Ntt Information Sharing Platform Lab., Musashino-shi, Jpn) NAKATSURU T (Ntt Information Sharing Platform Lab., Musashino-shi, Jpn)
Journal Title;NTT Tech Rev
Journal Code:F0282C
ISSN:1348-3447
VOL.2;NO.2;PAGE.77-84(2004)
Figure&Table&Reference;FIG.9, TBL.1, REF.7
Pub. Country;Japan
Language;English
Abstract;This paper describes a method for making personalized smart suggestions, in which context-adjusted and personalized activity-support information is provided. This method provides functions for evaluating how the current contextual information affects the user's activity and determining the best time for providing activity-support information to each user. We propose a general measure "motivation-indicator" for user-oriented context evaluation (e.g., level of burden, profit, and familiarity). We also discuss a self-tuning mechanism for personalization rules by analyzing how the behavioral response to the changing contextual information changes over time. The motivation-indicator provides a framework for analyzing whether or not activity-support information should be provided to a user in a specific context, which dispenses with the annoying procedure of pre-defining information-provision rules for each user. Thus our proposal has a beneficial effect on sharing context-based activity-support information among users without losing the personalization policy of each user. We have confirmed the feasibility of our method with a prototype. (author abst.)