Post on 07-May-2015
description
Introduction MyHealthEducator Conclusions
MyHealthEducatorPersonalization in the Age of Health 2.0
Luis Fernandez-Luque
Northern Research Institute, Tromso (Norway)
Adaptation and Personalization for Web 2.0, UMAP 2009, June2009
Introduction MyHealthEducator Conclusions
INTRODUCTION
Health Information and Personalization
Internet Health InformationMost of the Europeans use Internet to find health information.The Information Overload is growing (size and complexity).There are many initiatives to certify webs (e.g. HON).Resources can be both very good and dangerous.
Tailored Health EducationDesigned for specific health problems, mainly messages tomodify behaviors.Techniques: rule-base, questionnaires, predefined set ofresources, NLP, etc.
Introduction MyHealthEducator Conclusions
INTRODUCTION
The e-Patients in the Web 2.0
1 Socializing/Virtual Communities2 Creating and sharing information3 Accesing e-Health Services: such as Personal Health Records
(PHR) for controlling our personal information and an ecosystemof eHealth applications
Introduction MyHealthEducator Conclusions
MYHEALTHEDUCATOR
MyHealthEducator
What is MyHealthEducator?
It is a PhD Project: An adaptive Recommender System of healtheducation based on heterogeneous and changing data sources
A tool to assist the users to find relevant resources
Adaptable to changes in the user’s health
Enabling technologies1 Semantic-enhanced dynamic modeling2 Context-aware recommender techniques: user’s health is part of the
changing user context3 Others: Collaborative Techniques, Personal Health Records, analysis of
User Generated Content, Consumer Health Vocabulary, etc.
Introduction MyHealthEducator Conclusions
MYHEALTHEDUCATOR
Overview
Introduction MyHealthEducator Conclusions
MYHEALTHEDUCATOR
Health Repository
Introduction MyHealthEducator Conclusions
MYHEALTHEDUCATOR
User Modeling
Introduction MyHealthEducator Conclusions
MYHEALTHEDUCATOR
Recommender Engine and Evaluation
It is a hybrid Recommender System based on:Context-aware: semantic similarity to match videos with patient’shealth contextCollaborative filtering: highly rated videos by similar users (e.g.health, age, gender)
Different configurations will be tested (prefiltering with collaborativetechniques, weighted combination, etc.) in real setting:
1 Beta application (web-gadget) application installable in PHRs2 Integrated and tested in ongoing research projects (e.g. Tromso
Telemedicine Lab projects, European Projects)3 Evaluation will be focused on the technical validation, not a clinical
study.
Introduction MyHealthEducator Conclusions
CONCLUSIONS
Conclusions
Work done so farStudy of the literature: Personalization and Medicine 2.0,Tailored health education, Recommender Systems (accepted toMIE 2009)Study of User Generated Content: survey to e-Patients, analysisYouTube comments in health videos (presented at Medicine 2.02008, accepted to MIE 2009)Design of the system and prototype for crawling videos.
Future WorkDesign and first prototypes by the end of summer 2009Deployment and evaluation in 2010
Introduction MyHealthEducator Conclusions
CONCLUSIONS
Thanks for your attention
Luis Fernandez (luis.luque@norut.no)