Univerzální doporučovací systém
Thesis title in Czech: | Univerzální doporučovací systém |
---|---|
Thesis title in English: | Universal Recommender System |
Key words: | Doporučovací systémy, Internetové obchodování, Doménová nezávislost, Knowledge-based doporučování, Kolaborativní filtrování |
English key words: | Recommender Systems, E-commerce, Domain-Independence, Knowledge-Based Recommender, Collaborative Filtering |
Academic year of topic announcement: | 2009/2010 |
Thesis type: | diploma thesis |
Thesis language: | angličtina |
Department: | Department of Software Engineering (32-KSI) |
Supervisor: | prof. RNDr. Peter Vojtáš, DrSc. |
Author: | hidden - assigned and confirmed by the Study Dept. |
Date of registration: | 23.04.2010 |
Date of assignment: | 23.04.2010 |
Date and time of defence: | 30.05.2011 00:00 |
Date of electronic submission: | 14.04.2011 |
Date of submission of printed version: | 15.04.2011 |
Date of proceeded defence: | 30.05.2011 |
Opponents: | Mgr. Jan Dědek, Ph.D. |
Guidelines |
The goal of the thesis is to describe possibilities of creating a recommender system that is independent of the domain it is working on.
The first task is to define some abstract relationships that can be used for recommending. The recommender then utilizes the relationships for making recommendations using various approaches, recommending should be independent of the decision taking time (number of sessions). The adaptation to an existing system will be done by mapping the existing relationships to the abstract ones. The final goal is to implement an experimental prototype of such a recommender and adapt it to at least three existing systems. The assignment will be specified in more detail continuously. |
References |
Jerome Kunegis, Alan Said, Winfried Umbrath: The Universal Recommender, September 2009
http://adsabs.harvard.edu/abs/2009arXiv0909.3472K Geoffrey J. Hueter, Steven C. Quandt, Noble H. Hueter: Universal system and method for representing and predicting human behavior patent description http://www.freepatentsonline.com/20090248599.pdf J. Ben Schafer, Dan Frankowski, Jon Herlocker, and Shilad Sen: Collaborative filtering recommender systems http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.130.4520&rep=rep1&type=pdf Robin Burke: Hybrid Web Recommender Systems http://josquin.cs.depaul.edu/twiki/pub/Csc594/ArticleList/burke-hybrid.pdf and some more if needed. |
Preliminary scope of work |
Doporučovací systémy jsou programy, které se danému uživateli pokouší nabízet objekty (např. knihy, hudbu, novinové články), které by pro něj mohly být zajímavé. Tyto systémy mohou být užitečné pro uživatele, který se rozoduje mezi mnoha takovými objekty a nechce se zabývat procházením informací o všech dostupných objektech.
Tradiční doporučovací systémy jsou specifické pro danou doménu nabízených objektů. Náš pohled je obecnější, zahrnuje doménovou nezávislost. Cílem práce je vytvořit prototyp doporučovacího systému, aplikovatelného na různé domény. Po přizpůsobení by takový doporučovací systém byl schopen doporučovat objekty z domény. |
Preliminary scope of work in English |
Recommender systems are programs that try to present items such as books, music, news, etc. that are likely to be interesting for a given user. These systems may be helpful for users that are choosing between a large number of items and aren't willing to browse information about all available items.
Traditional recommender systems are specific to a particular domain of recommended items. Our view is more general and involves domain independence. The thesis aims to create a prototype of a recommender system that would be applicable to various domains. After some adaptation it would be able to recommend the items of the domain. |