Re: PageRanking with DIH
: I have computed pagerank offline for document set dump. I ideally : want to use pagerank and solr relevency score together in formula to : sort search solr result. I have already looked at : http://wiki.apache.org/solr/SolrRelevancyFAQ#How_can_I_increase_the_score_for_specific_documents : and found that indextimeboost is useful. I want to know how can I use : indextimeboost ? i would strongly suggest thta instead of using index time boost you use a boost function on a numeric field (the very next section of that SolrRelevancyFAQ) I've updated the page to try and make this alternative method more obvious, and mentioned the use of ExternalFileField (for the case where you want to be able to update these rankings w/o reindexing) http://wiki.apache.org/solr/SolrRelevancyFAQ#How_can_I_increase_the_score_for_specific_documents http://wiki.apache.org/solr/SolrRelevancyFAQ#How_can_I_change_the_score_of_a_document_based_on_the_.2Avalue.2A_of_a_field_.28say.2C_.22popularity.22.29 -Hoss
PageRanking with DIH
Hi, I have indexed documents and computed pagerank for documents. I want to update pagerank for indexed document and sort solr search result with pagerank. I did some research and found that index time boost can be used, but I don't know how to use it. Can I boost document at index time with DIH ? Can anybody help me in this regard ? Can I use Solr relevancy score with PageRanking score to sort search result? Any suggestions are welcome !! Thanks
Re: PageRanking with DIH
On 12 June 2012 13:04, vineet yadav vineet.yadav.i...@gmail.com wrote: Hi, I have indexed documents and computed pagerank for documents. I want to update pagerank for indexed document and sort solr search result with pagerank. Your question is not entirely clear: What is pagerank in this case? A custom score that you can compute at indexing time, and by which you want to order retrieved results? If so, just add a pagerank field to your Solr records, ignore Solr's order, and instead sort results by that field. I did some research and found that index time boost can be used, but I don't know how to use it. Can I boost document at index time with DIH ? Can anybody help me in this regard ? Can I use Solr relevancy score with PageRanking score to sort search result? Any suggestions are welcome This is confused: Do you want your pagerank as the sole basis for the ranking of returned results, or do you want it to be one of multiple (weighted) criteria? Maybe you should read http://wiki.apache.org/solr/SolrRelevancyFAQ Regards, Gora
Re: PageRanking with DIH
Hi Gora, Thanks for reply. I have computed pagerank offline for document set dump. I ideally want to use pagerank and solr relevency score together in formula to sort search solr result. I have already looked at http://wiki.apache.org/solr/SolrRelevancyFAQ#How_can_I_increase_the_score_for_specific_documents and found that indextimeboost is useful. I want to know how can I use indextimeboost ? Thanks Vineet Yadav On Tue, Jun 12, 2012 at 1:32 PM, Gora Mohanty g...@mimirtech.com wrote: On 12 June 2012 13:04, vineet yadav vineet.yadav.i...@gmail.com wrote: Hi, I have indexed documents and computed pagerank for documents. I want to update pagerank for indexed document and sort solr search result with pagerank. Your question is not entirely clear: What is pagerank in this case? A custom score that you can compute at indexing time, and by which you want to order retrieved results? If so, just add a pagerank field to your Solr records, ignore Solr's order, and instead sort results by that field. I did some research and found that index time boost can be used, but I don't know how to use it. Can I boost document at index time with DIH ? Can anybody help me in this regard ? Can I use Solr relevancy score with PageRanking score to sort search result? Any suggestions are welcome This is confused: Do you want your pagerank as the sole basis for the ranking of returned results, or do you want it to be one of multiple (weighted) criteria? Maybe you should read http://wiki.apache.org/solr/SolrRelevancyFAQ Regards, Gora
Re: PageRanking with DIH
On 12 June 2012 13:51, vineet yadav vineet.yadav.i...@gmail.com wrote: Hi Gora, Thanks for reply. I have computed pagerank offline for document set dump. I ideally want to use pagerank and solr relevency score together in formula to sort search solr result. I have already looked at http://wiki.apache.org/solr/SolrRelevancyFAQ#How_can_I_increase_the_score_for_specific_documents and found that indextimeboost is useful. I want to know how can I use indextimeboost ? [...] That depends on how you are indexing data into Solr. That page explains how to do index-time boosting at record level, and field level for XML documents uploaded to Solr with post.sh. If you are using the Solr DataImportHandler, you can boost records, but not individual fields, as far as I am aware. Please take a look at this thread for an example: http://lucene.472066.n3.nabble.com/Index-time-boosting-with-DIH-td3206271.html It would help if you did some basic groundwork, tried out things for yourselves, and asked more specific questions. You might wish to read http://wiki.apache.org/solr/UsingMailingLists Regards, Gora
RE: PageRanking with DIH
To boost a document with DIH, see this section about $docBoost in the wiki here: http://wiki.apache.org/solr/DataImportHandler#Special_Commands. If you're using a RDBMS for source data, your query would have something like this in it: select PAGE_RANK as '$docBoost', ... from ... etc If you don't want to boost entire documents but have it be very flexible at query time, see the page on Extended Dismax, especially the boost function section: http://wiki.apache.org/solr/ExtendedDisMax?highlight=%28edismax%29#bf_.28Boost_Function.2C_additive.29 . Also, the Packt Solr book (SmileyPugh) has a nice section about boosting scores based on page-rank or popularity type fields. In the old first edition its chapter 5, enhanced searching. James Dyer E-Commerce Systems Ingram Content Group (615) 213-4311 -Original Message- From: Gora Mohanty [mailto:g...@mimirtech.com] Sent: Tuesday, June 12, 2012 4:54 AM To: solr-user@lucene.apache.org Subject: Re: PageRanking with DIH On 12 June 2012 13:51, vineet yadav vineet.yadav.i...@gmail.com wrote: Hi Gora, Thanks for reply. I have computed pagerank offline for document set dump. I ideally want to use pagerank and solr relevency score together in formula to sort search solr result. I have already looked at http://wiki.apache.org/solr/SolrRelevancyFAQ#How_can_I_increase_the_score_for_specific_documents and found that indextimeboost is useful. I want to know how can I use indextimeboost ? [...] That depends on how you are indexing data into Solr. That page explains how to do index-time boosting at record level, and field level for XML documents uploaded to Solr with post.sh. If you are using the Solr DataImportHandler, you can boost records, but not individual fields, as far as I am aware. Please take a look at this thread for an example: http://lucene.472066.n3.nabble.com/Index-time-boosting-with-DIH-td3206271.html It would help if you did some basic groundwork, tried out things for yourselves, and asked more specific questions. You might wish to read http://wiki.apache.org/solr/UsingMailingLists Regards, Gora
