XML Retrieval dokumen bahasa indonesia menggunakan algoritme simnomerge

Abstract

This research is comparing the retrieval of relevant documents to irrelevant documents by using 130 documents of medicinal plants and 20 queries. Simnomerge algorithm is used in this research because different XML contexts are kept separate for the weighting. Each query is divided into 5 queries for unstructured retrieval system, 8 queries for XML retrieval with XML context depth 1, and 7 queries for XML retrieval with XML context depth 2. The result of this research shows that the XML retrieval can improve performance of the retrieval system by getting more relevant documents. At the relevant documents, average precision of XML retrievals query for an XML context depth 1 is 0.911, or a 15.5% increase from the unstructured retrieval system. Average precision of XML retrievals query for XML context depth 2 is 0.981, or a 22.5% increase from the unstructured retrieval system. Meanwhile, for the irrelevant document the average precision of XML retrieval for XML context depth 1 is 0.246 or a 36.5% decrease from the unstructured retrieval system. The average precision of XML retrieval for XML context depth 2 is 0.329 or a 28.8% decrease from the unstructured retrieval system.