Information Studies

M. A. Angrosh ¹, Shalini R. Urs ²

Affiliations
1 Indian Institute of Plantation Management, Bangalore, IN
2 Information and Communication Division and Professor, Department of Library and Information Science, University of Mysore, Mysore, IN

Abstract

Knowledge Organization Systems (KOS), are essential prerequisites for efficiently managing the rapidly increasing digital data repositories, besides their being vital components of digital libraries. However, various inadequacies associated with traditional digital content management mechanisms in achieving the required degree of control and freedom over digital data has hindered the information community in obtaining optimal benefits from available digital resources. Knowledge representation is at the heart of the entire system and has crucial significance for its contribution in organizing and managing digital information. While 'dictionaries', the initial mode of knowledge representation lacked the primary binding relationships between terms and corresponding definitions, 'thesaurus', the next immediate form of knowledge representation could only add synonyms and antonyms to the existing collection of terms and definitions without addressing the problem of relationships. Other forms of knowledge representation, such as, taxonomies brought in the required etymological information, by providing for hierarchical arrangement of synonyms and antonyms. While taxonomies have the ability in creating a tree structure of entities and in deriving association relations between classes, such techniques have limitations in providing the required degree of relatedness. Currently, ontologies, applied in knowledge representation, attempt to bridge existing gaps, primarily by providing for higher degrees of associative relationships between concepts could pave the way for designing semantic web-based information services. During the last few years, metadata (data about data) has played a significant role in information management systems. The primary objective of metadata has been to facilitate an end-user in finding the required piece of information. However, the absence of contextual relevance in these data sets has resulted in questioning the value of metadata in content management. The presence of contextual information in metadata, pertinent to digital content being captured can aid to a large extent in designing effective information services. Knowledge representation systems such as ontologies can be utilized for representing domain specific knowledge characteristics, which could be utilized for deriving 'semantic metadata', i.e. those forms of metadata that capture contextually relevant information about digital content. Such contextually relevant metadata that would form valuable knowledge structures can be exploited for retrieving actionable information. The present paper intends to draw upon the value of semantic metadata derived through a domain specific ontology, which would be tapped by information retrieval mechanisms, to provide search strategies to yield higher precision rates. The paper draws upon the development of an agri-pest domain specific ontology for modeling the agriculture-pest-disease-pesticide domain specific knowledge, which would be utilized for mapping content related data with domain knowledge. The resulting knowledge base will be used for information retrieval through contextually relevant entry points.

Full Text