Problem statement: Given a directed social graph, have to predict missing links to recommend users (Link Prediction in the graph) Data Overview Taken data from facebook's recruiting challenge on Kaggle: Kaggle Data contains two columns source and destination each edge in the graph - Data columns (total 2 columns): • source_node int64 • destination_node int64 Mapping the problem into supervised learning problem: Generated training samples of good and bad links from the given directed graph and for each link got some features like no. of followers, is he followed back, page rank, Katz score, Adar index, some svd features of adj matrix, some weight features, etc. and trained ML model based on these features to predict link. Some reference papers and videos: • Cornell • lichtenwalter2010 • YouTube Business objectives and constraints: • No low-latency requirement. • The probability of prediction is useful to recommend the highest probability links. The performance metric for supervised learning: • Both precision and recall are important so an F1 score is a good choice • Confusion matrix