Link Prediction: Missing Edges and Probabilistic Methods

This lecture by the instructor covers the topic of link prediction in undirected networks, focusing on the prediction of missing edges. Various methods such as Jaccard coefficient and logistic regression classifiers are discussed for link prediction. The lecture also delves into probabilistic methods for classification, including logistic regression with latent variables. Additionally, the lecture explores the challenges of causal inference under network interference, emphasizing the estimation of treatment effects in the presence of network dependencies.