際際滷

際際滷Share a Scribd company logo
0息 LexisNexis
QUINN Query updates for news monitoring
1息 LexisNexis
QUINN Query updates for news monitoring
Objectives
 Information specialists build complex Boolean queries for news retrieval
 News topics evolve over time: new entities/terms/concepts/events
 Search strings are long and complex: difficult to see how query can be
improved
 Question: can we assist users with query term suggestions?
 Hypothesis: result sets contain potentially relevant new query terms
 Term extraction technology implemented for the COMMIT/SWELL
project
First results
 Dedicated UI for term suggestion, showing:
- original query
- documents from news stream
- list of suggested query terms
 Evaluated with 6 users from 2 organizations so far:
- users recognize potential
- noisy term suggestions need improvement
- additional wishes: suggestions for Boolean operators, wildcards
 User experiment finalised later this week

More Related Content

Commit Community Days 2015: QUINN valorisation project

  • 1. 0息 LexisNexis QUINN Query updates for news monitoring
  • 2. 1息 LexisNexis QUINN Query updates for news monitoring
  • 3. Objectives Information specialists build complex Boolean queries for news retrieval News topics evolve over time: new entities/terms/concepts/events Search strings are long and complex: difficult to see how query can be improved Question: can we assist users with query term suggestions? Hypothesis: result sets contain potentially relevant new query terms Term extraction technology implemented for the COMMIT/SWELL project First results Dedicated UI for term suggestion, showing: - original query - documents from news stream - list of suggested query terms Evaluated with 6 users from 2 organizations so far: - users recognize potential - noisy term suggestions need improvement - additional wishes: suggestions for Boolean operators, wildcards User experiment finalised later this week