Abstract:
Semantic Web, as an extension of the traditional web, is concerned about the vast amount of unstructured data, and with its motive to make the entire knowledge content machine-readable, as well as machine-interpretable, all the processes of structuring the data are highly significant. Knowledge representation in trees has been a familiar mechanism for some time. However, such representations lack in existence when it comes to document content. This thesis properly presents a general mechanism that can generate a representation of the concepts of a document in the form of the knowledge tree. This rooted tree helps represent the contents of a document in an organized way as well as to find the core concepts of the document. We more considerably augment knowledge from various knowledge bases and analyze those data by mapping it with an existing ontology to obtain the taxonomy. We explain how this can be effective to create hierarchical concept recommendations of a document as well as to categorize documents easily. Finally, we introduce a framework for multilingual and able ontology to adopt new languages, also the addition of new data to the existing sources. The framework enriches the domain of the current ontology by integrating an infinite number of languages through mapping the dictionaries. Hence, the framework helps make the whole system and the central knowledge repository language independent. To conclude, we present the results obtained by the experimental implementation of the frameworks to demonstrate the accuracy of the tree and concept hierarchy to amply fulfill our ultimate goal.
