Semantic Information Retrieval Model by Spectral Clustering
Annie Jones · International Journal of Advanced Research in Computer Science · 2014
The Web which is increasing day by day has huge volume of unstructured data, with several aims, qualities and aspects which makes retrieval a tedious task. Semantic web which extend our current web has a focus to retrieve the data more precisely with vocabularies. These vocabularies are understood by the people and computer. Ontology, the core concept of semantic web which explodes data from knowledge base consists of instance of classes. This paper proposes method on how retrieval of information semantically can be done from heterogeneous data store. Here clustering methodology is used to match both ontology and the user query. The algorithm will navigate into the deep roots of the ontology structure and group the similar nodes with the query. The collected similar data are stored in a buffer area to produce an optimized output. The clustering of the data is done semantically to achieve higher relevancy. The spectral clustering algorithm which is used to achieve clustering semantically will locate the sparsely located data and match them efficiently. The basic idea is to collate the web not only to link the large heterogeneous documents but also to instruct meaning of the information in those documents.