A Keyword Recommendation System for GenBank

Abstract

With the progress of Human Genome Project, a huge amount of genomic data is stored in various databases and is open to the public through Internet. The size of these databases grows rapidly and, for example, the size of GenBank1, which is a well-known DNA sequence database, reaches more than 5 million entries. However, as the size of database grows, it becomes more difficult for users to retrieve necessary data. Many of genome databases consist of full text data and the keyword search method is widely used for their retrieval. Hence, when a user submits a simple query with a general keyword, he receives a large number of search results and is obliged to add another keywords to reduce the number of results. In this paper, we propose a keyword recommendation system[1, 2] that recommends proper keywords to the user and show a prototype where the technique is applied for GenBank. Advantages of our system are as follows.