Pembuatan Result Snippet pada Mesin Pencari Berbahasa Indonesia dengan Menggunakan Pseudo-Relevance Feedback

Abstract

A search result snippet is a quick excerpt from a document retrieved by a retrieval system (search engine) that explains the existence of the keywords entered by a user. Snippets generated by a search engine must be able to provide enough information so that a user can make decision on whether or not a document is relevant based on his/her information needs. This study implements pseudo-relevance feedback techniques for selecting the sentences to be displayed as a snippet and evaluates the accuracy of the snippet generated by the system. Application of pseudo-relevance feedback in this study resulted in an accuracy of 90.71% from the documents that are considered relevant.