Metode Pemilihan Fitur Dokumen Bahasa Indonesia yang Terkelompok pada Mesin Pencari

Abstract

The large amount of information particularly in the form of large quantities of documents will required a large amount of time and effort to search if done manually. On a vector space, documents are represented by terms. More terms mean higher-dimensional data which makes search more difficult to perform. A large number of documents affects the performance of the search engine to return the documents that are relevant to the user's desires. This study implements correlation coefficient method and compareit with the chi-square method. In this study different levels of accuracy are produced. Correlation coefficient method has an accuracy of 68% while the chi-square method produced an accuracy of 58%