Comp 336 & 529
Big Data Analytics
大数据分析作业代写 From the degree distribution point of view, the network should be between scale-free and random with sublinear preferential ……
1.I chose option 2 in this assignment, specifically, the ego 0 in the Facebook data is used. I’ve eliminated all disconnected parts from the node list and edge list. There are altogether 324 nodes and 5028 edges in this dataset, which matches the requirement of this assignment.
2.In this part, I chose the Networkx framework to visualize the network. In this part, I chose the kamada_kawai_layout among all 7 different layouts provided by NetworkX because it showed the basic structure information without mass 大数据分析作业代写
up with the complex edges. Each node in this network has a size proportionality to the degree of the node. So the nodes in the center of the graph are larger while there are small nodes with little connections scattered in the corner of the graph. Also, I adjusted the line width and line color so that they didn’t overlap with each other too much.
3.The degrees of each node is calculated and the estimated probability is shown below:
4.From the degree distribution point of view, the network should be between scale-free and random with sublinear preferential attachment. I have tried to fit the degree distribution curve and search for the Gama value. The result is shown below:
and the fitted gama=0.872 with an estimated covariance equal to 0.00011. Based on the summary on the 54th page in the course slice, the almost linear degree distribution indicates that the network should be a Scale-Free network. However, since the number of nodes is too small when the degree is large enough, the variance of the estimated density on the right hand of the distribution is larger. Based on this observation, if we adjust the weights in the fit procedure, the curve is more like a Random network.
更多代写：
发表回复
要发表评论，您必须先登录。