https://www.aminer.org/citation

Using NetworkX (https://networkx.github.io/) for network analysis (creating network, reporting network statis- tics, calculating three node centrality measures, and implement two link prediction methods)

1. Select data & create network define what kind of network I would like to analyze. The options including: 1) treat papers as nodes and citation relationships as edges, 2) treat authors as nodes, and co-author relationship as edges, 3) treat authors as nodes, and citation relationship as edges.

2. Network analysis & basic statistics In this part, I'm report the basic network statistics include: 1) num- ber of nodes in the network 2) number of edges in the network 3)average degree of node 4) radius of the network 5) diameter of the network 6) density of the network. Ref: https://networkx.github.io/documentation/stable/auto_ examples/basic/plot_properties.html#sphx-glr-autoexamples-basic-plot- properties-py. Next, draw a figure representing the node degree distribution in the network.

3. Node Centrailty Analysis Pick three centrality metrics that I'm interested to investigate and report at least three findings (e.g., what are the nodes with high centrality values, how the centrality values distributed in the network) Ref. https://networkx.github. io/documentation/stable/reference/algorithms/centrality.html

4. Link prediction Pick one unsupervised and one supervised link prediction algorithm, imple- ment the algorithms, and compare the performance of the selected two ap- proaches. You must consider at least one ranking based evaluation metric and one classi cation-based evaluation metric.