PERBANDINGAN METODE CLUSTERING MENGGUNAKAN HIERARCHICAL CLUSTERING DAN PARTITIONAL CLUSTERING UNTUK MENGELOMPOKKAN DOKUMEN BERITA
| Main Authors: | Kusumawardani, Yunita, Hamzah, Amir, -, Suraya |
|---|---|
| Format: | Article info application/pdf eJournal |
| Bahasa: | eng |
| Terbitan: |
Jurusan Informatika INSTITUT SAINS & TEKNOLOGI AKPRIND YOGYAKARTA
, 2018
|
| Subjects: | |
| Online Access: |
http://journal.akprind.ac.id/index.php/script/article/view/711 http://journal.akprind.ac.id/index.php/script/article/view/711/542 |
Daftar Isi:
- The development of the internet in Indonesia increasingly advanced and sophisticated. Similarly mass media that has been a lot of venturing into the media online to facilitate users in accessing the news or information via online. And here began of growing and abudant news or information in the text. So make the news has not categorized properly. Then cause difficulties in searching appropriate news categories and should spend a lot of time in searching the abundant news text on the internet. Research to classsifying this news of the text, using 2 clustering method that is Hierarchical Clustering by using Agglomerative Hierarchical Clustering and Partitional Clustering method using K-Means Clustering. Two clustering methods are used, with the aim of knowing and comparing which methods are better and resulting in higher accuracy in grouping of news texts, as well as faster time in grouping processing. This research use data totaling 30 news data. From 30 news data is divided into 8 news data for news document collection 1, 10 news data for news document collection 2, and 12 news data for news document collection 3. The result of research using agglomerative hierarchical clustering obtained accuracy value on news document collection 1 that is 62 , 5%, news document collection 2 is 50%, news document collection 3 is 33,33%, for k-means clustering is obtained result of highest accuracy value in news document collection 1 is 62,5%, news document collection 2 is 60 %, News document collection 3 is 33,33%. While the average time obtained with agglomerative hierarchical clustering is 29,5 s and k-means clustering is 30,19 s.