International Conference on Advances in Computer and Information Technology - ACIT 2013
Author(s) : AINI RACHMANIA , JAFREEZAL JAAFAR
With the Internet fast progressing, the number of news documents increases significantly and the demand for easy navigation around the news becomes pivotal. The online news domain needs a classifier that is able to classify news articles accurately at a low cost computation. Apart from that, because it is always updated, the classifier also needs to be able to detect new topic. This paper presents the topic identification and category classification method which enable categorization of news articles and identification of topic as new news arise. During the training phase, keywords are extracted from each document. Then, a document is classified to a predefined category. Lastly, the topic of the document is identified. Testing was conducted on Indonesian news corpus. The result shows that the classifier was able to classified Indonesian text documents with a satisfactory accuracy level as high as 93.84%.