[ANSWERED] In this project, you will be learning from a training network and trying to predict whether edges…

In this project, you will be learning from a training network and trying to predict whether edges exist among test node pairs. The training network is a fragment of the academic co-authorship hypergraph. The nodes in the network- -authors-have been given randomly assigned IDs, and each undirected hyperedge corresponds to a paper which links the two or more authors of the paper.’ You will have to decide how best to represent the graph, and how to use this in a predictive model. The training network reflects the co-authorship of papers up to a specific point in time, focussing on individuals in a specific academic subcommunity. The testing network captures new co-authorships that have arisen more recently, i.e., subsequent to the time when the training graph was constructed. The test data is a list of 2000 undirected edges, and your task is to predict if each of those test edges are really part of the authorship network or if they are fake. Half of the test edges are real, while the other half do not actually exist. Note that test edges are supplied as undirected edges, that is, they comprise a pair of authors, which differs from the training graph format, which is in the form of hype edges. You must predict whether these test pair of authors will coauthor at least one paper together in the ‘future’, potentially in collaboration with other authors. To make the project fun, we will run it as a Kaggle in-class competition. Your assessment will be partially based on your final ranking in the privately-held competition, partially based on your absolute performance and partially based on your report.
Attachments:
6387215_1_a1-1.docx