Mesin Pencari SMS berbasis Android dengan Menggunakan Metode N-gram

Abstract

Search engines Short-Messaging Service (SMS) on mobile devices are still using the exact match method. This method can only find SMS containing words, which one exact matches of the query. Therefore, a method to correct the spelling of the text is needed. Trigram similarity measure method is one of spelling correction methods developed which gives a relatively good performance. Therefore, this study applied the method of N-gram with bigram models and trigram models for SMS in Indonesian. The result of this research is a search engine with an average precision (AVP) 49.6% for bigram models and 80.8% for trigram models.