The estimation uses the MLE method described by (2) and implemented by (https://github.com/ericsuh/dirichlet).
The input csv should be comma separated with no spaces and a header row.
All ids should be numeric and start at 0.
The strength of a node's community membership (member_prob) should be a real number
between 0 and 1.
The format is:
node_id,community_id,member_prob
If the csv containing the community data is examples/example.csv, then the usage is:
python communityprior.py examples/example.csv alpha.csv beta.csv
Two files containing the priors, alpha.csv and beta.csv, will be written to the current directory.
- Zhang, H., Qiu, B., Giles, C. L., Foley, H. C., & Yen, J. (2007, May). An LDA-based community structure discovery approach for large-scale social networks. In Intelligence and Security Informatics, 2007 IEEE (pp. 200-207). IEEE.
- Minka, T. (2000). Estimating a Dirichlet distribution.