Your SlideShare is downloading. ×
Adapting Alax Solr to Compare different sets of documents - Joan Codina
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×

Saving this for later?

Get the SlideShare app to save on your phone or tablet. Read anywhere, anytime - even offline.

Text the download link to your phone

Standard text messaging rates apply

Adapting Alax Solr to Compare different sets of documents - Joan Codina

710
views

Published on

See conference video - http://www.lucidimagination.com/devzone/events/conferences/ApacheLuceneEurocon2011 …

See conference video - http://www.lucidimagination.com/devzone/events/conferences/ApacheLuceneEurocon2011

One of the main features of Solr is Faceted Search. Facets are the top terms present in the results of a query. But facets do not indicate the most statistically relevant terms of a query, that is, these terms that are more present in the documents selected by the query than in the rest of the collection. A critical factor in making such statistical insights broadly useful is to make them visual -- i.e., using charts and graphs that display these quantitative relationships. We will present how to adapt Ajax-Solr to find the most prominent terms of a query compared to the full set or just another query. We are going to present and example on how this can be used to find current topics in the news, and extract that information into visually communicative charts and graphs.

Published in: Technology, Education

0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
710
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
9
Comments
0
Likes
0
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. Adapting Ajax Solr to Compare different sets of documents Joan Codina-Filba, Barcelona Media-Innovation Center 19-Oct-2011
  • 2. What I Will Cover More than just Facets We do research on opinion mining To Give an overview of thousands of opinions After analyzing the contents of them: • How to explore the key concepts expressed in the opinions. • Are the key concepts the most common ones, or the most differentiating? Use of Solr and Ajax Solr for a more intelligent data navigation 2
  • 3. My Background Joan Codina Filbà Barcelona Media – Innovation Center Senior Researcher Experience on • Data => Text => Web => Opinion MINING • Integration of NLP tools. • Multimodal search • Some teaching 3
  • 4. The Challenge Companies want to know what does the web 2.0 say about them: Too many writers, in many places, not enough time to read. – Detect hot topics (Quality, Price, service?) – Detect strengths / weakness – Knowledge about the market Use NLP to detect key concepts in text User generated content, noisy, irony, dispersion From data to information and knowledge 4
  • 5. Use Case Example Analyze the users in MySpace Combination of demographic data and analyzed text Possibility to "navigate" the data to get “knowledge” of the users. Exploring thousands of messages, before sending them to a “blind” , “automatic” text mining tool to find relevant aspects. An “OLAP” for data and text, a kind cube view of the opinions 5
  • 6. Solr Facets Simple statistic... – For fields with a single value: SELECT field, count(x) FROM table GROUP BY field • Simple aggregation statistic – For text fields ? Facets vs IDF – The most common facets give less information – The rare ones, are not useful 6
  • 7. Information given by Facets8 1 Facets 0,97 idf information 0,86 0,75 0,64 0,5 0,43 0,32 0,21 0,10 0 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 7
  • 8. Information given by Facets8 1 Facets 0,97 idf information 0,86 0,75 0,64 Too Too 0,5 rare Common 0,43 0,32 0,21 0,10 0 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 8
  • 9. The Information is in the “not so common" terms Doing a search and looking at facets to characterize the data: – For most of the queries the most common facets do not change significantly – Sorting by frequency: • The order almost is the same. • Difficult to see the changes The information is not on the terms frequency but in the relative changes on these frequencies 9
  • 10. Example Exploring MySpace posts Demographic Data: Age, Gender, Country Text Content 10
  • 11. Example After a query (country = MX) 11
  • 12. Example After a query (country = MX) Find the 10 differences !! 12
  • 13. Example After a query (gender= Male) 13
  • 14. Example After a query (gender= Male) Find the 10 differences !! 14
  • 15. The Information is on the differences What do we want to know: – Which are the terms that use the Males and/or people from Mexico ? – OR – Which are the terms that better identify or differentiate them? 15
  • 16. Example After a query (country = MX) Substract them !! 16
  • 17. Example  After a query (country = MX)Faceted Results: The most used words by the mexicansStatistical Results: The words that differentiate the mexicans from the rest of the world 17
  • 18. Example  After a query (gender = M)Faceted Results: The words that Males use mostStatistical Results: The words that differentiate the males from the females 18
  • 19. The information is on the difference We can compare a subset with the full set: But we can compare different subsets: 19
  • 20. The information is on the difference We can compare by gender, doing different queries 20
  • 21. Data Charts We can apply differences to charts 21
  • 22. Data Charts We can apply differences to charts: And show the difference with the global set 22
  • 23. Data Charts Data Charts can show the differences Males Females 23
  • 24. Data Charts Even through queries Males Females 24
  • 25. Ajax Solr  Single Query / Resultset Query AjaxInterface SOLR Resultset User Browser Javascript front end Server 25
  • 26. Ajax Solr Extension  Multiple Queries / Resultsets Query AjaxInterface SOLR Resultset User Browser Javascript front end Server 26
  • 27. Ajax Solr Extension: Multiple Sets Each set has its own query – A query is composed of different query terms – The terms that define the set, can not be deleted – Query terms can be copied/deleted through the different sets Widgets can view/compare data from the different result sets. 27
  • 28. Ajax Solr Extension: Multiple Sets Set 0 Independent widgets Base Data Set 1 Set 2Widgets related to this set Widgets related to this set Own Filter Query Own Filter Query 28
  • 29. Conclusions Facets give basic statistics The information is on the differences Ajax Solr can be adapted to manage different sets of data Charts and graphics help us to have an overview of the data: information. 29
  • 30. Sources The code is not ready to be published • Was done for SolrJs • Migrating to AjaxSolr (Lucene-Solr 4.0) But it will be... Links – http://www.barcelonamedia.org/personal/joan.co dina/en 30