| Institution: | Slovak University of Technology |
| Technologies used: | Ruby |
| Inputs: | programmers' acitivity - web browsing, code review etc. |
| Outputs: | recommended sources for information |
Addressed problem
Recommender systems are an integral part of modern-adaptive web nowadays. The need for the personalized web increases day by day, while users are generally overwhelmed by the amount of experienced information. This phenomenon can be seen in every domain where users interact with the information sources (web, program codes etc.). Personalised recommendation is the most used approach to satisfy both-users and business (in the per-formance mean) respectively.
Description
We propose novel method which extends the task of group recommendation to the single-user recommendation. The aggregation of single-user profiles in order to obtain one group profile combines users' preferences and also in some settings introduces items variety, which can be interesting from the recommendation improvement point of view. The main difference between classic single-user collaborative recommendation process and proposed approach is that we recommend not based on the user to user similarity, but based on the similarity between users and a virtual group to which is the programmer assigned (represented as the virtual user). Thanks to various settings as group size or inner-group similarity it is possible to control and improve results in order to fulfil specific goal – to obtain various results, or to focus on specific interest area. Proposed approach consists of three basic steps:
- virtual groups construction and virtual user preferences aggregation;
- similarity computation between virtual users and real users outside the group;
- generation of recommendation for specific user.
References
Kompan, Michal: Exploring Group Recommendation for Single-User Recommendation Tasks.
In: Student Research Conference 2012. Vol. 1 : 8th Student Research Conference in Informatics and Information Technologies Bratislava, April 25, 2012 Proceedings. - Bratislava : Nakladateľstvo STU, 2012. - ISBN 978-80-227-3689-3. - S. 217-224
Kompan, Michal – Bieliková, Mária: Personalized Recommendation for Individual Users Based on the Group Recommendation Principles. 2013. In: Studies in Informatics and Control. – ISSN 1220-1766. s. 11 (to appear)