OWLIM Overview

From Saltlux

OWLIM is a topic based search service started in 2008. Literally OWLIM is a combination of meaning “to put information together (jeongbo reul owluda)” in Korean language and “OWL” the semantic language.

Background of Birth of OWLIM
Current information search is basically focused to efficiently discover the information from unstructured data. Essentially because of unstructured form of data it might be a reasonable way to approach. And because of weak development of document describing technology information in comparison to target information, information search upto the present point is basically in the status level of frequency and accuracy oriented access. In this situation Objectives of OWLIM service are to classify unstructured documents in topic orientation and to test and experience of services that discover the relationship between documents.


Issues of existing search
Text analysis: It may occur not to define the concepts of homonymic, Meta language, and thesaurus documentations in the existing search. The concerned issue could have been resolved through separate after processing as needed.
Analysis of inter-relationship between indexing and query: Ranking problem created based on weight appraisal by measuring similarity between document and word might be raised when user mentioned word does not clearly stated.
Query analysis: Accurate expression of user’s needs is not easy and takes longer time. To hidden type expression users, a tool that supports concrete queries is desperately needed. However, what the current user can experience is a lower level approach like with search language recommendation.


Solution suggested by OWLIM
Topic clustering: OWLIM provides solution to narrowing user query ranges by analysis of total documents of search results and through classification of important keywords existing in target query related documents.
Visualization of information: OWLIM also provides a solution that gives instinctive sense in a short time for user’s information process by understanding quantity and procedure of documentation classified with topic clustering, forming topic map through measuring distance and size.