rSpringRank implements a collection of regularized, convex models (+solvers) that allow the inference of hierarchical structure in a directed network, which exists due to dominance, social status, or prestige. Specifically, this work leverages the time-varying structure and/or the node metadata present in the data set.
This is the software repository behind the paper:
-
Tzu-Chi Yen and Stephen Becker, Regularized methods for efficient ranking in networks, in preparation.
-
For full documentation, please visit this site.
-
For general Q&A, ideas, or other things, please visit Discussions.
-
For software-related bugs, issues, or suggestions, please use Issues.
rSpringRank is available on PyPI:
pip install rSpringRank
The dependencies needed to use graph-tool-based input are not installed by default. You will need to install:
- Graph-tool: https://graph-tool.skewed.de/
In your Python console, rSpringRank is invoked using:
>>> from reg_sr import rSpringRank, PhDExchange
>>> pde = PhDExchange()
>>> g = pde.get_data(goi="c18basic")
>>> rsp = rSpringRank(method="annotated")
>>> result = rsp.fit(g, alpha=1, lambd=1, printEvery=0) # actual computation; takes ~5 seconds
>>> pde.compute_basic_stats(rsp.sslc, primal_s=result["primal"])
>>> pde.print_sorted_mean(5, precision=3) # output the higher mean SpringRank categories
Group: 15; Mean: 0.171
Group: na; Mean: -0.043
Group: 22; Mean: -0.043
Group: 18; Mean: -0.043
Group: 21; Mean: -0.043As you noticed, most groups have the same mean SpringRank. This is the regularizer in effect. To plot the distribution, do this.
>>> pde.plot_hist(bin_count=20, legend=True)
TODO.
TODO.
TODO.
rSpringRank is open-source and licensed under the GNU Lesser General Public License v3.0.