Implementation of Data Mining Clustering Using the K-Medoids Method in Grouping Library Books Politeknik Negeri Balikpapan

Abstract

Politeknik Negeri Balikpapan’s Library, annually adds to its collection of reading books. In order to procure the book according to the needs of users, namely students, information on the collection of books needed or in demand is needed. To answer this problem, it is necessary to have a clustering system for existing books in the library by looking at the aspect of borrowing frequency. The Clustering system is made using the K-Medoids method with the selection of 3 Clusters, namely, very attractive, desirable and less desirable. From the results of data processing through the Rapid Miner Application with K = 3 , the results obtained are cluster_0 (low) consisting of 97 book titles with the frequency of borrowing books in the rare category in other words less desirable to borrow, cluster_1 (high) consists of 2 book titles which are the most in-demand group of books namely Teknologi Bahan dan Teori dan Praktik Hotel Front Office. Cluster_2 (medium) consisting of 8 titles which are books with moderate borrowing frequenc.  The results of the performance obtained Davies Bouldin index value of 0.287. The results of grouping the data of these books can be used as input for library managers in procuring book collections based on the frequency of borrowing books.