Friday, already! Happy fragile ceasefire in Gaza, happy wedding to Daniel and Anna, happy Martin Luther King Day, and happy presidential inauguration.
I pass by this mural of MLK almost every day, and I find it incredibly expressive. Maybe it has something to do with the Seattle weather. Today I thought he was looking tired, but hopeful.
Is there some research paper out there to distract me from thinking about politics too much? How about a networks paper with applications to politics? An Arxiv paper from the Physics and Society section was updated recently, and it caught my eye with some beautiful pictures. I’ll explain a little, or you can head straight for their figure on the interactions between presidential candidates and corporations.
In Googling hidden interactions: Web search engine based weighted network construction, Lee and co-authors look at complete graphs on politicians, physicists, and baseball players, where each edge is weighted by how many search results Google reported on a query representing both of its endpoints.
I’m not saying that this is a great idea (Google result counts are probably not that accurate), and I’m not saying that I would do it just like they did if it was my project (weighting edges with the raw count seems less interesting than the Jaccard coefficient). And I’m not saying that the name they invented for a maximum 1-out is better than the existing terminology. But I do love the graphic in figure 7! And some of the others are appealing, too.
(In further defense of the name “maximum 1-out”, as an established alternative to “maximum relatedness subnetwork” (MRS), it would probably be interesting to look at the maximum 2-out in these networks, as well. Good terminology makes research questions easier to come up with.)
Healthy Algorithms 