Fig. 2
From: Trust and class aware service discovery with dual control in the social Internet of Things
Conceptual and empirical validation of the proposed class–based greedy service discovery algorithm. (a) Conceptual structure of class ratings and trust filtering: for each query class \(\zeta _q\), neighbors with class rating \(CR(\zeta _q) \ge \alpha\) are retained, and the Top–K highest-ranked among them are selected for expansion. This pruning eliminates weak or redundant links and preserves only trustworthy, high-value connections in the ISWSIoT. (b) Empirical traversal trace on the Caenorhabditis graph (\(\alpha {=}0.6\), \(K{=}6\), \(H{=}6\)). The requester is shown in red, the discovered service provider in green, intermediate expanded nodes in blue, and the final path in orange. Grey edges indicate links pruned by \(\alpha\) filtering and Top–K control. The trace demonstrates bounded greedy traversal that successfully reaches the target in 4 hops, validating the practical effect of the conceptual design in (a).
