Table 6 Greedy mode statistical summary with 95% confidence intervals. The table reports the mean and 95% confidence interval [lower, upper] for success rate, average hops, average visited nodes, and latency (in milliseconds) across \(\alpha\) thresholds for each dataset. Each value is computed from \(N = 100\) independent discovery trials using bootstrap resampling (\(B = 1000\)) to quantify sampling variability. The visited nodes metric captures traversal overhead and exhibits wider confidence bounds on larger graphs (e.g., Epinions), reflecting greater variability in exploration paths across heterogeneous trust topologies. Success rates remain consistently high across datasets, while hops and latency exhibit narrow confidence bounds, confirming stable performance and reproducibility of the greedy discovery mechanism. The inclusion of uncertainty estimates for all metrics ensures that reported performance differences are statistically interpretable and not attributable to random variation.

From: Trust and class aware service discovery with dual control in the social Internet of Things