Keywords Recommender for Scientific Papers Using Semantic Relatedness and Associative Neural Network
F Nugroho (abc)
a) Master of Information Systems, Faculty Of Post Graduate
b) Universitas Komputer Indonesia
Jl. Dipati Ukur No.112-116, 40132, Bandung, Jawa Barat, Indonesia
c) fajarnugroho.id[at]gmail.com
Abstract
Keywords are important because they represent a simple form of content and have important role in finding articles in information gathering as well as effective indexing. Keywords are common in scientific papers; website (meta tag), news, etc. In this research, we are focusing on scientific writings. Scientific writings are also categorized as structured because they have firm parts such as title, abstract, content, and conclusion so whenever we are going to find important information, we can find them through specific parts such as title, abstract, and conclusion.Automatic keywords extraction from scientific papers is used to assist on suggesting keywords for indexers or to produce summary in a form of keywords for scientific writings.By using trained auto associative artificial neural network with chosen keywords data sets we are able to find keywords features in scientific documents and a combination of semantic approaches in which relation among words is an important aspect aside of words frequency to discover important things in a document, so we expect to produce recommended relevant keywords. So the automated keywords extraction and pre-included keywords from the writer will be compared to choose either the most relevant or complementary and providing recommendation or perhaps the extracted automated keywords and the keywords from the writer are exactly the same.
Keywords: Automatic Keyword Extraction (AKE); Semantic Relatedness; Neural Network
Topic: Informatic and Information System