This is a mass batch-processing task, rather than a search task. Mahout is the right Apache project for implementing this. It would then create a set of (document->document list). You could then add this to a Solr index. (And invert the graph and add those lists.)
It might be possible to do this with the code that is already available in Mahout. On Mon, Aug 30, 2010 at 6:24 AM, Ukyo Virgden <ukyovirg...@gmail.com> wrote: > Hi, > > Is there any implementation in solr or lucene for "affinity ranking"? I've > been doing some research for content based ranking models and came across > the paper "Improving search results using affinity Graph" > http://research.microsoft.com/apps/pubs/default.aspx?id=67818 > > Any thoughts? > > Cheers > Ukyo > -- Lance Norskog goks...@gmail.com
