Method for ontology representation from unstructured text corpus

Social media platforms has facilitated sharing the important significant events from individual’s daily lives. Users often communicate unconventionally, employing abbreviations and misspelled words. Consequently, it becomes imperative to pay close attention to unstructured text, which encompasses events from diverse sources such as social media posts, webpages and articles. It is equally essential to utilize ontology as a means of representing and storing knowledge. Visualization of the concepts and relations within the ontology allows for a clear and semantic representation of the data. This study centers on developing a method for constructing structured ontology representation from unstructured event-related news texts collected from diverse sources. The key idea is a phased approach combining NLP techniques and rule-based extraction to derive event schema, followed by ontology formalization using. Moreover, it offers improved outcomes for entity extraction, particularly regarding precision, recall and F1 score evaluation, as well as the quality of the derived schema.