Penggunaan Operator Belief Revision pada Temu Kembali Dokumen Bahasa Indonesia Model Boolean

Abstract

Retrieval results of conventional Boolean Model are based only on the exact matching between document and query without ranking. Extended Boolean Model has developed a search process using partial matching and ranking, but it is difficult to be applied for complex queries structures. Belief Revision model was introduced to improve the information retrieved by Boolean Models. The objective of this research is to implement Extended Boolean Model (EBM) and Belief Revision (BR) to rank Indonesian documents. Query and document are weighted in binary. Then, the similarity is calculated to obtain the rank. Information retrieved using EBM equals to BR with average precision of 0.5148