Identification of Successful Sellers in Online Auction
Ladislav Beránek, Václav Nýdl, Radim Remeš

Language: en
Last modified: 2015-03-13


The design of efficient methods concerning prediction of preferences of sellers on online auctions is an important problem in the study of Internet auctions. In this paper, a new prediction method is proposed based on Dempster - Shafer theory of evidence. The proposed method is based on the approach used in complex networks when determining influential nodes using for example centrality measures. The suggested method takes into account the degree and strength of each node  which is expressed as the number of positive evaluations (degree of reputation of a seller) and the price of sold goods (by certain seller). The effects of both reputation and price of each seller who sold certain goods are represented by basic belief assignments (BBAs). The proposed prediction of the choice of sellers on online auctions is then determined by a combination of these BBAs when so-called auction preference index is calculated.  Experiments are used to illustrate the effectiveness of the proposed method.


Dempster - Shafer theory; Online auction; Preference prediction; Influential sellers; Link prediction

Full Text: PDF