Friday, December 28, 2012

Social H-Index

Happy New Year, Everyone!

I remember the STOC/FOCS bibliography that David Johnson compiled in mid 90's (pre-web).  It had a lot of interesting "top-k" lists. One list remained on my mind: authors with most number of single-authored STOC/FOCS papers. Andy Yao topped that list,  far above the rest.

How does one measure the impact of an author or a publication? The h-index aims to improve over raw counts of papers or citations by combining both, but considers authors in isolation and does not consider collaborations. My coauthors and I have been playing with extensions that "socialize" the h-index. The main idea is, a paper by authors (a,b) should count to the social score of a if it is one of the top papers of b. Now, what is a top paper, how to extend this to multiple authors, etc. are addressed in our paper that also has empirical results contrasting the solo h-index vs social h-index.



Anonymous Anonymous said...

I believe your notion of social h-index is biased against theoreticians since (I would guess) theoreticians have relatively less coauthors as compared to other areas of computer science.

4:54 PM  
Anonymous Anonymous said...

There is some correlation between social h-index and the no of papers or no of coauthors, but it is not a perfect mapping.

-- Metoo

4:59 AM  

Post a Comment

<< Home