Figure 2
From: On the complexity of quantum link prediction in complex networks
Probability of sampling a useful link in QLP. Simulating QLP on a range of datasets from real-world complex networks52,53,54,55,56 we computed the total probability of obtaining a useful sample, \(p_\text {G}(t)\). In the first plot, for all network sizes, the probability saturates in the 0.1 to 1.0 range for increasing t, indicating it does not decrease with N. This is more explicit in the scatter plot of \(p_\text {G}\) vs N for different values of the time t of the quantum walk. In the third plot we note there is a tendency for the useful link probability to increase with the average connectivity of the network, at least for low values of \(k_\text {av}\). Overall, these results indicate the probability of sampling a useful link will typically be a constant overhead in the algorithm as N increases.
