Direct Term Feedback pada Temu-Kembali Informasi Bahasa Indonesia Menggunakan Model Bahasa

Abstract

The large amount of information has triggered the development of search engines to help the users in finding the required information. To retrieve information which is relevant to the users’ needs, the queries should be formulated correctly. Direct term feedback is a method that can help a user to directly judge the relevance of individual terms without interaction with feedback documents, taking a full control of the whole query expansion process. This technique uses Pseudo-relevance feedback which takes the top n-ranked documents and takes the top x-ranked terms from the relevant documents to get term feedback for users’ judgement. The purpose of this research is to implement query expansion with direct term feedback for information retrieval in Indonesian language with language models. This research uses 2095 documents and 20 testing queries. The result shows that direct term feedback is a good method to help the users in finding their required information. The performance of direct term feedback for information retrieval with language models gives a good result with around 35% of AVP value.