Semi-Automatic Domain Ontology Graph Generation System in Punjabi
Rajveer Kaur, Saurabh Sharma · 2016
Ontology is the presently rising research topic in the field of artificial intelligence, semantic web, and natural language processing. Manual ontology development is very complex and time consuming task. This research proposes and implemented a Semi-Automatic Punjabi Domain Ontology graph generation system which creates a framework semi-automatically known as Ontology Graph. It is the first time that this system has been developed for Punjabi language. Punjabi is the official language of Punjab state of India. The proposed system includes process of learning ontology that illustrate the way of learning and creates ontology graphs semi automatically from Punjabi text documents. Ontology learning process includes two phases namely Preprocessing phase and processing phase. Preprocessing phase is a representation of the Punjabi text in structured way. The proposed system starts by Pre-processing phase which includes Term extraction by applying input restrictions, removing special symbols, duplicate terms, stop words, and matching input terms with dictionary and gazetteer lists. Processing phase includes identify relationship between term and class, relationship between different terms in a class. The mapping of relationship between term and class is done according to weighted mapping. Terms that have frequency of occurrence greater than 3 are used to map the relationship between term and class. The relationship between different terms in class finds out the direct relation between terms in a class.