Abstract
Impact factor is a widely accepted quantitatively reputation model for literatures and authors. However, it cannot satisfy the requirements for evaluating literatures more exactly. In fact, there are much more information which can be included into value evaluation. The paper models the relationships among literatures, its author, periodicals and readers which can be regarded as recommendations for each other that are in some ways similar to hyperlinks in Google. We build a similar reputation model and propose a Page Rank-like algorithm. Comment is a key factor in reputation model that can be gained readily from forum systems. A reputation evaluation algorithm and a modification version, which can run with much less cost, are proposed. By simulations, we prove that the basic algorithm is feasible and convergent and the latter can mitigate time cost greatly with the loss of accuracy about 30%. Periodically computing the reputations using the basic algorithm is necessary. The new reputation model can be applied in academic forums and improve the academic evaluating mechanism.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have