Open AccessCommunity structure-aware fairness and goodness algorithm for link weight prediction
Author(s) -
Hiba Rashid Atiya,
Huda Naji Nawaf
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1804/1/012080
Subject(s) - node (physics) , goodness of fit , computer science , enhanced data rates for gsm evolution , representation (politics) , data mining , algorithm , machine learning , artificial intelligence , law , structural engineering , politics , political science , engineering
In this research, the problem of predicting the edge weight in the Bitcoin network has been resolved by utilizing the structure of the community based Fairness and Goodness. Community detection using Newman-Girvan algorithm has been applied to obtaine the trusted communities which depends on the features representation generally. The former may represent the trusted transactions that have trust value more that certain threshold. Concering the missing edge weight, the prediction is based on the fairness and goodness that supported by the above communities. In fairness measure, can capture how fair the node is in rating other nodes’ trust that has transactions (trusted transactions) with. While the goodness of a node shows how much this node is liked or trusted by other nodes that have transactions (trusted transactions) with. In both cases, only the nodes that are within community of the target node contribute with setting the fairness or the goodness of the target. The model was evaluated in practice using two real-world datasets; the Bitcoin-OTC and the Bitcoin-Alpha datasets. The experimental findings explain the effectiveness of the proposed comparable with other methods. The percentage error minimization is 18% for the Alpha dataset and 26% for OTC dataset.