EKSTRAKSI INFORMASI HALAMAN WEB MENGGUNAKAN PENDEKATAN BOOTSTRAPPING PADA ONTOLOGY-BASED INFORMATION EXTRACTION (OBIE

Information extraction is identification of structured information from unstructured text. Several approaches for information extraction had been developed by many researchers. This study covers semi-automatic extraction to identify NER (Named Entity Recognition) in natural language text. This research proposed a method that combined Ontology-Based Information Extraction (OBIE) with bootstrapping approach to extract web pages. The ontology was used to guide information extraction process by referring to the existing ontology concept. Bootstrapping was used to learn new facts from unlabeled text using a small number of label data (seed). A case study to apply this approach used dataset \"LonelyPlanet\" (Cimiano dkk., 2005). The performance evaluation achieved were 73% precison, 62% recall and 67% F-measure. Keywords: information extraction, NER, Named Entity Recognition, ontology, Ontology-Based Information Extraction, OBIE, bootstrapping, performanc