<![CDATA[News is a source of information that displayed to the general public about an event and presented in various languages. Usually, a website only allows user to search only in one language. This causes problems for users who want to find broader information more quickly in several languages ​​at once. These problems can be overcome by developing a cross language information retrieval system. The system can improve the time efficiency because it can return documents in two languages ​​by simply entering a query in one language only. One of the method that can be used to develop the system is BM25F method that can return relevant documents and handle structured documents. The news data structure used in training and testing is the title and the content part of the news. The data used in this study are 300 Indonesian news documents and 300 English news documents that will be used to test the boost value, the Indonesian queries, and the English queries. For the boost value testing, the highest precision@k value obtained when the title boost is 5 and the content boost is 1. This value will be used for query testing. Query testing is performed using precision@k and got the highest value of 0.98 when k=5 in the Indonesian queries test which returned Indonesian and English documents.]]>
Copyrights © 2020
