Thomson Scientific: Rescuing “lost” data

Industry : Publishing

Using Text Mining to apply thesaurus-based Indexing to digitized Print Material

Customer challenge

..............................................................................

Thomson Scientic had 49 bound volumes of Biological Abstracts® for 1926 to 1968 and needed to make the data searchable with the BIOSIS controlled vocabulary to offer a new web product to its clients.

Solution

..............................................................................

The approach

• Use entity extraction to obtain candidate terms from the titles and abstracts
• Map the extracted entities to the BIOSIS vocabulary
• Output the resulting indexing
• Publish the new product

Key ITM Benefits

..............................................................................

Top Line Results

• Project completed to schedule (5 months)
• 1.9M indexed records
• All records received at least required minimal indexing
• Throughput: 500ms per item

Biosis product presentation (253.75 ko)

Thomson project presentation (1.52 Mo)

Partner

..............................................................................

Intelligent Topic Manager (ITM) from MONDECA for

• Biosis thesaurus management (more than 2 millions concepts in 6 thesaurus)
• Normalize and publish content metadata

Luxid from TEMIS for entities extraction on digitized contents