Information Retrieval dengan Menggunakan Metode Latent Semantik Indexing (LSI) pada Proses Searching dan Klasifikasi Buku: Studi Kasus Perpustakaan STMIK Bumigora

Keywords: Information Retrieval, Singular Value Decompostion, Latent Semantic Indexing, term document matrix


A library in STMIK Bumigora as support unit in educational institution. It is provided to support teaching learning process. Enhencing library function optimally will improve the quality of education. The problems of STMIK Institution library are that in book classification by librarian in STMIK Bumigora is done manually, it means the books added and labelled to the shelves based on the information title as administrator knowledge, that sometimes books with the same tittle numbers are placed in different places. As a result, the visitor have problems when they search the book, sometimes book and the place is not as expected, for example the title and contents does not match. Therefore the research obataining IR of books or documents indexing in library can be derived by using Laten Semantic Indexing (LSI)method in searching process and clasifiying process. LSI represents terms of the document in the form of vectors which arrange in Term Document Matrix. The exellence of LSI is to deal with the polysemy and synonym words which usually found in documents. It works by counting the rapport among terms in one document with terms in other documents. It can be done by decomposing term document matrix as inverted file of LSI in order to get the rapport of those terms. So the relevant document though it does not contain any term of a query, stillit can be generated. Decomposing matrix can use Singular Value Decomposition (SVD) method. This research discusses IR of library system by appliying LSI method and using SVD as the decomposition method classification. Experiment has been conducted toward the document of book summary contained on back cover. The preliminary data is in the form of scanned textbooks in Bahasa, then it is converted into text document files (.txt file). This proposed method is expected to assist the process of classifying and searching books in the library. LSI can provide appropriate information to its visitors to find books that is relevant to the query inputs.


[1] Eko Prasetyo.2012.“Data Mining Konsep dan Aplikasi Menggunakan Matlab”. Yogyakarta: ANDI.
[2] Jianxiong Yang, Junzo Watada, “Decomposition of Term-Document Matrix Representasion for Clustering Analysis”, (IEEE International Conference on Fuzzy Systems )2011, p978
[3] Michael W. Berry.Susan T. Dumais.Gavin W. O’Brien.1995. Society for Industrial and Applied Mathematics.SIAM review.”Using Linear Algebra for Intelligent Information Retrieval”.Journal Jstor. PP. 573-995.
[4] Riyanarto Sarno. Yeni Anistyasari. Rahimi Fitri. 2012.“Semantic Search”. Yogyakarta: ANDI.
[5] Wei Song.Soon Cheol Park. 2009. Division of Electronics and Information Engineering, Chonbuk National University, Jeonju, 561756, Republic of Korea .“Genetic algorithm for text clustering based on latent semantic indexing”. Journal of Machine Learning Research. Elseiver.Vol.57. Issue 11-12Yates, RB. “Modern Information Retrieval”. Boston, USA. Addison WesleyPearson International Edition 1999