2 quick ideas: 1) The nouns and verbs have to be in separate fields. So, you would make two analysis stacks. One only stores terms marked as 'noun' and the other only stores terms marked as 'verb'. 2) Store all terms in one field. Create a Similarity class that scores from the payloads. (There is a method for this, but I remember a problem where it is not called. Don't know if this is fixed yet.)
On Tue, May 8, 2012 at 4:36 AM, hemantverm...@gmail.com <hemantverm...@gmail.com> wrote: > Hi All > > I have intergrated UIMA and Solr and successfully indexed the documents. > My requirement is to boost nouns over verbs. (Lets take for example field : > description) > Below is the sample index contents. > > - <doc> > - <arr name="concept"> > <str>Water Curses</str> > <str>Prospect Hummer</str> > </arr> > <str name="description">people are different in nature some bahaves better > but some behaves like animal.</str> > <str name="id">3</str> > <str name="language">english</str> > - <arr name="sentence"> > <str>people are different in nature some bahaves better but some behaves > like animal.</str> > </arr> > - <arr name="tag"> > <str>nns</str> > <str>ber</str> > <str>jj</str> > <str>in</str> > <str>nn</str> > <str>rb</str> > <str>vbz</str> > <str>jjr</str> > <str>cc</str> > <str>dti</str> > <str>vbz</str> > <str>cs</str> > <str>nn</str> > <str>.</str> > </arr> > <str name="title">say hello to everybody</str> > > Can anyone suggest what are the next steps to fulfill this requirement? > What I got in index is sequential posTag in different field and I still not > got any clue how to index *word*_posTag so that this *word*_NOUN can be > boosted more than only *word*_VERB in search query. > > What I am thinking, is right approach or not? > Or is there any better solution for boosting documents on the basis of > posTag? > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/SOLR-UIMA-posTag-tp3971101.html > Sent from the Solr - User mailing list archive at Nabble.com. -- Lance Norskog goks...@gmail.com
