<html><div style='background-color:'><DIV class=RTE>
<DIV></DIV>>We ought to do some testing to see whether our choice of BM25 parameters </DIV>
<DIV></DIV>>is good or not. I don't recall how we chose them originally - it may be
<DIV></DIV>>they came from Stephen Robertson, but it rather looks like we just used
<DIV></DIV>>the parameters which would give TradWeight, but with 0.5 for D which
<DIV></DIV>>gives the halfway point between BM11 and BM15.
<DIV></DIV>>
<DIV></DIV>>Andy MacFarlane did some evaluation work in the BrightStation days, but
<DIV></DIV>
<P>yes, but I didn't do any comparisons with other weighting schemes. I did do a lot of parameter tuning, but any conclusions I drew from it would not be that useful as these parameters vary from collection to collection.</P>
<P>>I don't recall if he got any results before they shut down development. </P>
<P>>Results with BM25 in other systems, are probably applicable if we can </P>
<DIV></DIV>>find some.
<DIV></DIV>
<DIV> </DIV>
<DIV>Comparison of results between BM25 and other weighting scheme goes on all the time at</DIV>
<DIV>TREC (see <A href="http://trec.nist.gov">http://trec.nist.gov</A>).</DIV>
<DIV> </DIV>
<DIV>cheers</DIV>
<DIV> </DIV>
<DIV>andy</DIV></div><br clear=all><hr>MSN Premium gives you PC protection, junk-mail filters, advanced communication tools and great software like MSN Encarta® Premium. <a href="http://g.msn.com/8HMAENUK/2749??PS=47575" target="_top">Click here for a FREE trial!</a> </html>