Applying Recent Enhancements in Reputation Based Mechanisms for推荐.pdfVIP

Applying Recent Enhancements in Reputation Based Mechanisms for Promoting Trust in Mobile Communities Marcus Dobrinkat Helsinki University of Technology mdobrink@cc.hut.fi Abstract They rely mostly on community users’ feedback and utilize rules and values of the specific community to maintain or- One inherent problem with virtual communities is how to der based on these values. Individuals that do not act ac- assess the trustworthiness of unknown peers. Reputation cording to the community rules are punished with negative based trust mechanisms are one means to help users decide feedback and will be stigmatized with a low reputation value whether to rely on a stranger or not. Using eBay’s feedback visible for the whole community, which effectively excludes forum as an example, we show how simple reputation based that peer from being addressed. trust mechanisms work. We evaluate shortcomings of these One of the first prominent real world examples of reputa- mechanisms and give an overview of reasonable enhance- tion mechanism was eBay’s feedback forum. Resnick and ments. Later we show how effectiveness is improved by var- Zeckhauser have done a statistical analysis of the reputation ious means including policy based trust measures and using and transaction data [2] and show that the system works more sophisticated models for calculating reputation values. although they reveal problems in the mechanism. They When applying the mechanisms to mobile communities, claim that eBay’s users trust in the system makes the syst