Group Recommender System for Restaurant Lunches

• Main Author: Hallström, Erik
• Format: Others
• Language: English
• Published: KTH, Skolan för datavetenskap och kommunikation (CSC) 2013
• Subjects: Computer Sciences; Datavetenskap (datalogi)
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-142393

A group recommender system for lunch restaurants is developed. The user interface is an Android application which is run on a smartphone. The system features a novel approach for implicit rating collection when a user browses a list of item descriptions. The individual recommendation is based on extracting and comparing tf-idf feature vectors of menu texts as well as individual rankings of the restaurants. The group recommender system works by aggregating the individual estimated scores of the members in the group, and in addition to this machine learning methods are used to capture the group dynamics in the group decision. === Inom ramen for examensarbetet utvecklas ett system for grupprekommendation av lunchrestauranger. Anvandarapplikationen laddas ner och kors pa en Android smartphone. Systemet anvander sig av en ny metod for att samla implicit information fran anvandarnas nyttjande av applikationen. Den individuella rekommendationen bygger pa att extrahera och jamfora tf-idf feature-vektorer fran menyernas text samt individuell rangordning av restaurangerna. Grupprekommendationen fas genom att aggregera de individuella estimerade betygen hos anvandarna i gruppen, och dartill anvands maskininlarningsmetoder for att simulera gruppdynamiken av gruppvalet.