Perolehan Dokumen dengan Metode Latent Semantic Indexing untuk Dokumen Bahasa Indonesia

Daniel Daniel, Viny Christanti M


The main focus of this study is to develop a system to retrieve document based on query. The document that used in this paper were collected from blogspot which contains of travelling and culinary topic. This system produce ranked relevant document based on similarity of each document using query. In this paper, we used two methods. The first method is LSI (Latent Semantic Indexing), in LSI we use Singular Value Decomposition (SVD) to reduce the rank of matrix which is too big for system to compute. The second method is VSM (Vector Space Model), this method is used to compute the weight of each word. The VSM was used because it was initially one of the best methods that used for retrieval. In this paper, we used precision and recall to measure the accuracy of both methods. It turns out that the precision and recall using LSI is 0.16569 and 0.25 while using VSM is 0.93961 and 0.81

Key words Indonesian blogspot, Information Retrieval, Latent Semantic Indexing, Vector Space Model.


  • There are currently no refbacks.