Ontology-Based Automatic Classification for News Articles in Indonesian Language

• Main Authors: Prajna Basnur, Dana Sensuse
• Format: Article
• Language: English
• Published: Universitas Indonesia 2010-10-01
• Series: Makara Journal of Technology
• Subjects: ontology, naïve-bayes, stopwords, stemming
• Online Access: http://journal.ui.ac.id/technology/journal/article/view/173

<p>Searching specific information&nbsp;will be difficult if only rely on query. Choosing less specific queries will result many irrelevant information fetched by&nbsp;the system. One of the most successful way to overcome this problem is to perform document classification based on&nbsp;the topic. There are many methods that can be used to conduct such a document classification, such as statistical and&nbsp;machine learning approaches. However, those document classification methods require training data or learning&nbsp;documents. In this study, the authors attempted to classify documents using a method that doesn’t require learning&nbsp;documents. This classification method uses ontology to classify documents. Document classification by using ontology&nbsp;is done by comparing the value of similarity among documents and existing node in the ontology. A document is&nbsp;classified into a category or a node, if it has the highest similarity value in one of the nodes in the ontology. The results&nbsp;showed that the ontology can be used to perform document classification. Recall value is 97.03%, precision is 91.63%,&nbsp;and f-measure is 94.02%.</p>