RE: dropping noise words and maintaining the relevancy
Hello, On dropping noise words we have scenario that we have to only drop ending noise words. For e.g. 160 Associates LP, the noise words here are Associates and LP but we only want to drop the LP one which is a ending noise word. If we use stop words, it will drop both words and make search key as 160. Any suggestion? Thanks in advance. -Original Message- From: Susheel Kumar [mailto:susheel.ku...@thedigitalgroup.net] Sent: Thursday, October 31, 2013 9:59 PM To: solr-user@lucene.apache.org Subject: RE: dropping noise words and maintaining the relevancy Thanks, Kranti. Nice suggestion. I'll try it out. -Original Message- From: Kranti Parisa [mailto:kranti.par...@gmail.com] Sent: Thursday, October 31, 2013 3:18 PM To: solr-user@lucene.apache.org Subject: Re: dropping noise words and maintaining the relevancy One possible approach is you can populate the titles in a field (say exactMatch) and point your search query to exactMatch:160 Associates LP OR text:160 Associates LP assuming that you have all the text populated into the field called text you can also use field level boosting with the above query, example exactMatch:160 Associates LP^10 OR text:160 Associates LP^5 Thanks, Kranti K. Parisa http://www.linkedin.com/in/krantiparisa On Thu, Oct 31, 2013 at 4:00 PM, Susheel Kumar susheel.ku...@thedigitalgroup.net wrote: Hello, We have a very particular requirement of dropping noise words (LP, LLP, LLC, Corp, Corporation, Inc, Incoporation, PA, Professional Association, Attorney at law, GP, General Partnership etc.) at the end of search key but maintaining the relevancy. For e.g. If user search for 160 Associates LP, we want search to return in their below relevancy order. Basically if exact / similar match is present, it comes first followed by other results. 160 Associates LP 160 Associates 160 Associates LLC 160 Associates LLLP 160 Hilton Associates If I handle this through Stop words then LP will get dropped from search key and then all results will come but exact match will be shown somewhere lower or deep. Regards and appreciate your help. Susheel
Re: dropping noise words and maintaining the relevancy
One possible approach is you can populate the titles in a field (say exactMatch) and point your search query to exactMatch:160 Associates LP OR text:160 Associates LP assuming that you have all the text populated into the field called text you can also use field level boosting with the above query, example exactMatch:160 Associates LP^10 OR text:160 Associates LP^5 Thanks, Kranti K. Parisa http://www.linkedin.com/in/krantiparisa On Thu, Oct 31, 2013 at 4:00 PM, Susheel Kumar susheel.ku...@thedigitalgroup.net wrote: Hello, We have a very particular requirement of dropping noise words (LP, LLP, LLC, Corp, Corporation, Inc, Incoporation, PA, Professional Association, Attorney at law, GP, General Partnership etc.) at the end of search key but maintaining the relevancy. For e.g. If user search for 160 Associates LP, we want search to return in their below relevancy order. Basically if exact / similar match is present, it comes first followed by other results. 160 Associates LP 160 Associates 160 Associates LLC 160 Associates LLLP 160 Hilton Associates If I handle this through Stop words then LP will get dropped from search key and then all results will come but exact match will be shown somewhere lower or deep. Regards and appreciate your help. Susheel
RE: dropping noise words and maintaining the relevancy
Thanks, Kranti. Nice suggestion. I'll try it out. -Original Message- From: Kranti Parisa [mailto:kranti.par...@gmail.com] Sent: Thursday, October 31, 2013 3:18 PM To: solr-user@lucene.apache.org Subject: Re: dropping noise words and maintaining the relevancy One possible approach is you can populate the titles in a field (say exactMatch) and point your search query to exactMatch:160 Associates LP OR text:160 Associates LP assuming that you have all the text populated into the field called text you can also use field level boosting with the above query, example exactMatch:160 Associates LP^10 OR text:160 Associates LP^5 Thanks, Kranti K. Parisa http://www.linkedin.com/in/krantiparisa On Thu, Oct 31, 2013 at 4:00 PM, Susheel Kumar susheel.ku...@thedigitalgroup.net wrote: Hello, We have a very particular requirement of dropping noise words (LP, LLP, LLC, Corp, Corporation, Inc, Incoporation, PA, Professional Association, Attorney at law, GP, General Partnership etc.) at the end of search key but maintaining the relevancy. For e.g. If user search for 160 Associates LP, we want search to return in their below relevancy order. Basically if exact / similar match is present, it comes first followed by other results. 160 Associates LP 160 Associates 160 Associates LLC 160 Associates LLLP 160 Hilton Associates If I handle this through Stop words then LP will get dropped from search key and then all results will come but exact match will be shown somewhere lower or deep. Regards and appreciate your help. Susheel
