Improving Lucene's geographical data search performance with R-trees Evgeny Shadchnev
This event took place on 24th August 2007 at 12:00pm (11:00 GMT)
Knowledge Media Institute, Berrill Building, The Open University, Milton Keynes, United Kingdom, MK7 6AA
Lucene, a state-of-the-art open source information retrieval library, is an efficient solution for indexing and searching textual data. However, some Lucene usage scenarios require handling of geographically augmented data, that is, text documents that contain geographical coordinates (e.g. wikipedia pages about cities). This data is best searched using spacial access methods, such as R-trees, provided that the number of unique documents is large enough to benefit from this approach. An extension to Lucene that improves its speed at searching geographically augmented data is described.
Lucene, a state-of-the-art open source information retrieval library, is an efficient solution for indexing and searching textual data. However, some Lucene usage scenarios require handling of geographically augmented data, that is, text documents that contain geographical coordinates (e.g. wikipedia pages about cities). This data is best searched using spacial access methods, such as R-trees, provided that the number of unique documents is large enough to benefit from this approach. An extension to Lucene that improves its speed at searching geographically augmented data is described.
Click below to play the event (30 minutes)