to ISI Home Page
isd home
About ISD
education at isd
employment
environment
news
people
research
AI Seminars
div3admin

environment
Andrew Philpot

Jose-Luis Ambite

Eduard Hovy


"Automated Retrieval, Analysis, and Classification of Internet Recipes for Tailored Food Pantry Distribution"   Small Research Awards Seminar

10/24/2003: 10:30am - 12:00pm
11th Floor Large Conference Room

Abstract: Participants: Andrew Philpot, Eduard Hovy and Jose-Luis Ambite. Participants at food pantry distribution sites increasingly often receive significant quantities of fresh produce as well as nonperishable staples and prepared foods. Some participants experience difficulty using this produce in a timely manner because of unfamiliarity with particular ingredients, limited food preparation skills, and/or lack of kitchen equipment. The Quick! Help for Meals (quickhelp.org) program was established to address these issues. A Quick! Help participant undergoes a tailored profiling session which assesses the preferences, needs, abilities, family setting, and skill level of the household's primary food preparer. The participant receives a personalized, customized recipe booklet generated on-site from the profile responses and an annotated internal recipe database. Quick! Help has been shown to improve participants' utilization of received produce and to increase their awareness of health benefits of eating fresh produce. (Quick! Help was developed in conjunction with Susan Evans and Peter Clarke of the USC Medical School, Institute of Health Promotion and Disease Prevention Research). One obstacle to wider usage of Quick! Help is the significant manual labor required to collect, analyze, and annotate recipes for inclusion in the recipe base. In this ISI Research Fund effort, the authors developed a system for collecting recipes from various internet resources containing a specified food ingredient. Retrieved recipes are analyzed for appropriateness, graded for level of difficulty, and integrated into the XML recipe base for use in the overall Quick! Help system. In this talk we will briefly highlight the Quick! Help system, describe our technical approach, show some examples of the results, and talk about future directions of this work.


Last updated: Mon Jun 19 17:44:06 2006

 

 

 

 

 
USC Home Page ISI Home Page