Link Prediction and Unlink Prediction on Dynamic Networks
Authors:
Christina Muro,
Boyu Li,
Kun He
Abstract:
Link prediction on dynamic networks has been extensively studied and widely applied in various applications. However, temporal unlink prediction, which also plays an important role in the evolution of social networks, has not been paid much attention. Accurately predicting the links and unlinks on the future network greatly contributes to the network analysis that uncovers more latent relations be…
▽ More
Link prediction on dynamic networks has been extensively studied and widely applied in various applications. However, temporal unlink prediction, which also plays an important role in the evolution of social networks, has not been paid much attention. Accurately predicting the links and unlinks on the future network greatly contributes to the network analysis that uncovers more latent relations between nodes. In this work, we assume that there are two kinds of relations between nodes, namely long-term relation and short-term relation, and we propose an effective algorithm called LULS for temporal link prediction and unlink prediction based on such relations. Specifically, for each snapshot of a dynamic network, LULS first collects higher-order structure as two topological matrices by applying short random walks. Then, LULS initializes and optimizes a global matrix and a sequence of temporary matrices for all the snapshots by using non-negative matrix factorization (NMF) based on the topological matrices, where the global matrix denotes long-term relation and the temporary matrices represent short-term relations of snapshots. Finally, LULS calculates the similarity matrix of the future snapshot and predicts the links and unlinks for the future network. Additionally, we further improve the prediction results by using graph regularization constraints to enhance the global matrix, resulting that the global matrix contains a wealth of topological information and temporal information. The conducted experiments on real-world networks illustrate that LULS outperforms other baselines for both link prediction and unlink prediction tasks.
△ Less
Submitted 16 December, 2021; v1 submitted 18 January, 2021;
originally announced January 2021.
Group Size Effect on the Success of Wolves Hunting
Authors:
Ramon Escobedo,
Denys Dutykh,
Cristina Muro,
Lee Spector,
Raymond Coppinger
Abstract:
Social foraging shows unexpected features such as the existence of a group size threshold to accomplish a successful hunt. Above this threshold, additional individuals do not increase the probability of capturing the prey. Recent direct observations of wolves in Yellowstone Park show that the group size threshold when hunting its most formidable prey, bison, is nearly three times greater than when…
▽ More
Social foraging shows unexpected features such as the existence of a group size threshold to accomplish a successful hunt. Above this threshold, additional individuals do not increase the probability of capturing the prey. Recent direct observations of wolves in Yellowstone Park show that the group size threshold when hunting its most formidable prey, bison, is nearly three times greater than when hunting elk, a prey that is considerably less challenging to capture than bison. These observations provide empirical support to a computational particle model of group hunting which was previously shown to be effective in explaining why hunting success peaks at apparently small pack sizes when hunting elk. The model is based on considering two critical distances between wolves and prey: the minimal safe distance at which wolves stand from the prey, and the avoidance distance at which wolves move away from each other when they approach the prey. The minimal safe distance is longer when the prey is more dangerous to hunt. We show that the model explains effectively that the group size threshold is greater when the minimal safe distance is longer. Although both distances are longer when the prey is more dangerous, they contribute oppositely to the value of the group size threshold: the group size threshold is smaller when the avoidance distance is longer. This unexpected mechanism gives rise to a global increase of the group size threshold when considering bison with respect to elk, but other prey more dangerous than elk can lead to specific critical distances that can give rise to the same group size threshold. Our results show that the computational model can guide further research on group size effects, suggesting that more experimental observations should be obtained for other kind of prey as e.g. moose.
△ Less
Submitted 4 August, 2015;
originally announced August 2015.
