Algoritma Term Frequency – Inverse Document Frequency (TF-IDF) dan K-Means Clustering Untuk Menentukan Kategori Dokumen

The development of technology is speedy; one of the results is developing documents in research articles. Searching for documents in a repository will take a long time if they are not stored grouped by document category. One way to define document categories is clustering. The usefulness of document clustering, to make it easier to find documents by certain categories. The clustering process uses the Term Frequency - Inverse Document Frequency (TF-IDF) algorithm and K-Means. TF-IDF is used to find document weights, while K-Means is for the clustering process. The test documents or dataset were grouped as many as 93 documents, with various themes and document contents. The K-Means cluster quality assessment process results using the Silhouette score; the optimal number of clusters is 4 clusters. This is obtained by looking at the fluctuation in cluster size and thickness of the silhouette plot.