Network Analysis of Word Vectors

This project is a part of semester credits in Social Network Analysis. As per project requirements, the main code and visualizations are in IPython notebooks. The gephi visualizations are in img directory. Listed below are the files in the repository with description

• report.pdf gives a brief overview of the project with results.
• network.gexf is the network file to be opened with gephi.
• Notebooks/createNetwork.ipynb : Use trained word embeddings to create a network.
• basic statistics.ipynb : Measures stats of the network eg. Degree Distribution, Centrality etc.
• network diffusion.ipynb : Network-based diffusion analysis (SIR, SIS, SI, etc.)
• CD.ipynb: Community Discovery. Includes Girvan-Newman, Louvain, Label Propagation, k-clique and Demon. Compares partition across them.

Requirements

The codes are heavily utilizing NDlib and networkX which can be installed using

$pip install ndlib networkx

Dataset

The dataset used for the creation of word-embedding is Sentiment140 dataset. Sentiment140 is a collection of 1.6 million tweets extracted using the Twitter API.