Pembobotan RIDF pada Mesin Pencari Bahasa Indonesia untuk Ekspansi Kueri Menggunakan Analisis Konteks Lokal

Abstract

Automatic query expansion is a technique for dealing with the fundamental issue of word mismatch between queries and documents in information retrieval. A number of approaches on query expansion have been studied, especially on techniques that analyze the corpus to discover word relationships (global technique) and those that analyze documents retrieved by initial query (local technique). Both techniques have their own advantages and limitations. Local context analysis is an automatic query expansion which is a combination of global and local techniques. The purpose of this research is to implement query expansion with local context analysis using TF (Term Frequency) RIDF (Residual Inverse Document Frequency) weighting method, a variation of IDF (Inverse Document Frequency) that assigns weights to terms according to the difference between the logs of the actual IDF and its prediction by Poisson model. The obtained performance of information retrieval was 15% average precision. The results also showed that the number of top-ranked documents and passages did not significantly affect the average precision. The more influential factor was the number of query expansions.