METODE WEIGHTED MAXIMUM CAPTURING UNTUK KLASTERISASI DOKUMEN BERBASIS FREQUENT ITEMSETS
| Main Authors: | Pradnyana, Gede Aditra; Jurusan Teknik Informatika, Fakultas Teknologi Informasi, Institut Teknologi Sepuluh Nopember Kampus ITS Keputih Sukolilo, Surabaya, Jawa Timur, Djunaidy, Arif; Jurusan Sistem Informasi, Fakultas Teknologi Informasi, Institut Teknologi Sepuluh Nopember Kampus ITS Keputih Sukolilo, Surabaya, Jawa Timur |
|---|---|
| Format: | Article application/pdf eJournal |
| Bahasa: | ind |
| Terbitan: |
Jurnal Ilmu Komputer
, 2013
|
| Subjects: | |
| Online Access: |
http://ojs.unud.ac.id/index.php/jik/article/view/8408 |
Daftar Isi:
- Document clustering based on frequent itemsets is one of the new document clustering method that can be usedto overcome the problem of high-dimensional space of the document being clustered. Maximum capturingtechnique is one document clustering algorithm based frequent itemsets that can generate better clusteringquality compared to those produced by other similar algorithms. The maximum capturing technique still has thelack or weakness, ie: not accounting for the weight of a word (item) in the calculation of frequent itemsets whenthe document similarity and the cluster formation process does not take into account the global information ofthe cluster previously formed. In this research developed a new method for clustering documents based frequentitemset namely weighted maximum capturing method (WMC), to correct deficiencies maksimum capturingaccuracy so that the quality of document clustering results can be improved. In the weighted maximumcapturing method, document similarity is computed by combining the cosine similarity method and Jaccardcoefficient based on the same number of frequent itemsets owned so that the weight of items in itemsets can betaken into account, while in the process of constructing cluster adapted single linkage agglomerativehierarchical clustering algorithm. Experimental results show the value of F-measure and purity of WMC methodis better than the earlier method, that is equal to 0.723 for the F-measure with improvement ratio of 2.8% and apurity value of 0.73 with improvement ratio 3.3%.