Internet Recommendations
| Project status | definitely finished |
| Project time | 01.01.2007- 31.12.2008 |
| Funding source | Land (außer NRW) |
| Keywords | Recommender Systems; Rating of Products; Group recommendations; Media |
Consumers nowadays are confronted with confusing variety of alternatives when choosing products and services. Especially with respect to hedonic products (like movies, music and games), which nearly have no search attributes and are difficult to evaluate before consumption, the consumers need high quality information to avoid wrong decisions. These information are scarce resources and therefore precious.
Web-based recommender systems pledge relief for the problem by analyzing product evaluations of consumers, applying mathematical algorithms and generating choice recommendations out of the evaluations. Existing recommender systems have essential limitations: On the one hand surveys show that existing recommender systems only model consumers preferences in an unsatisfying manner and many recommendations are error prone. On the other hand the applicability of conducted recommendations is rather low because recommender systems usually only provide recommendations for individual consumers - while in reality hedonistic products usually are consumed together with other individuals.
The objective of this research project is the development of a new-type recommender system that integrates results from consumer research as much as information technology driven group research to increase the forecast quality of existing systems and includes the fact of group consumption of hedonic products. The system should become developed and validated - at which for the first time knowledge about the