Sie sind auf Seite 1von 2

Using NetworkX write a program to implement the following.

A group of Three members can be formed and involve in solving assignment.

Assignments should be submitted in Jupyter notebook with relevant output and explanation.

Do not copy from peers and maintain academic integrity. Any defaulters found will be attract
penalty.

Hard deadline to submit assignment will be 15th November 2019.

Assignment-2

1. Explain how and why do communities form?

2. Write a short notes on each technique and Pick any two data sets from prominent datasets
available online and find communities using

a) Girvan and Newman algo(Edge Betweenness)

b) Clique Percolation Method

c) K-clique community

d) Label Propagation

e) Louvian method (Modularity based)

f) Spectral clustering/Eigen Vector Centrality

g) bruteforce technique (comeup with your own logic to find communities in a dataset)

3. Discuss different application of detecting communities like 1) multicasting a message of interest to


a community instead of sending into each one in the community.2) Recommending a set of products
to a community based on their interest.

Assignment-3

Theme based work based on topics of Interest like:

1. Make a jupyter notebook with relevant explanation on “Machine Learning for Dynamic Social
Network Analysis”------ (Bonus work) few links are shared below

2. Static Network Analysis(do not depend on time) :

What are communities at time T?

Who belong to a community ?

How closely knit is this community?

3. Temporal/ Evolution Analysis:

How did this community form?(Ex: Criminal Networks)


Which communities are stable?

Why did these communities form or dissolve?

4. Predictive Analyses:

Is this community likely to grow?

Will these nodes continue as a community in future?

Are dominant roles emerging in this community?

5. Information Cascading Techniques:

What is the role of Information cascading?

Describe Cascading behaviour in Networks?

Find a node through which Information cascading can be maximized?

How to become Viral in Social Networks?

Epidemics model of Information cascading like SIR and SIS.

6. Link Prediction:

Find the probability of link formation between any pair of chosen nodes in a network?

Apply any Machine learning based prediction model to predict friendship or chose a new
team of people from the given network?

Links for further study(if interested)

https://www.kdnuggets.com/tag/social-network-analysis

https://ieeexplore.ieee.org/abstract/document/5562752

https://dl.acm.org/citation.cfm?id=1557108

https://ieeexplore.ieee.org/abstract/document/7965209

https://link.springer.com/article/10.1007/s11432-014-5237-y

https://link.springer.com/chapter/10.1007/978-3-319-05579-4_45

https://dl.acm.org/citation.cfm?id=1864760

Das könnte Ihnen auch gefallen