Performance Assessment of Collaborative Trust Models

Abstract

This study assessed existing Collaborative Models for trust awareness to ensure effective friend-to-friend collaboration in an overlay network for better control of private data towards a robust social media interaction. The result obtained showed that the Trust-Aware Model (T-AM) demonstrated significant reliability regarding recommendation accuracy and convergence. This was because benchmarking the result from assessing the T-AM with the result from that of assessing existing collaborative models, showed that the T-AM had an accuracy of 0.875, and that of the existing collaborative model was 0.750. While the T-AM had a convergence value of 0.125, that of the existing collaborative model was 0.250. The existing model had an effectiveness value of 0.1875, while the T-AM had an effectiveness value of 0.1094. This implies that the T-AM provided improved recommendation accuracy considering the standard scale of 0 to 1, and convergence (in terms of time) on a scale of 0 to 1. The implication of this is that the T-AM’s ability to make recommendations is very significant since its accuracy value was 0.875 as compared to 0.750 of the existing collaborative models. For convergence, based on the scale rating provided above, it means that the T-AM provided accurate recommendations in a convergence of time of 0.125 as compared to 0.250 for the existing collaborative model. However, in terms of effectiveness, the T-AM performed less with an effectiveness value of 0.1094 as compared to the effectiveness value of 0.1875 of the existing collaborative models. The study concluded that trust data was effectively managed using the distributed hash table and symmetric replication methods, with significant improvement in reputational accuracy and convergence without compromising on scalability and secured online collaboration. Keywords: Trust Awareness, Effectiveness, Convergence, Accuracy, Collaborative Model