Penggalian Frequent Closed ltemsets Dengan Multiple Minimum Support

Purwanti, Endah (2009) Penggalian Frequent Closed ltemsets Dengan Multiple Minimum Support. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 5107201002-Master_thesis.pdf Restricted to Repository staff only Download (5MB)

Abstract

Penggalian frequent pattern memegang peranan penting dalam proses penggalian aturan asosiasi. Namun demikian, penggalian frequent pattern sering kali menghasilkan sejumlah besar frequent itemsets dan aturan, sehingga dapat mengurangi efisiensi dan keefektifan proses penggalian karena mengharuskan pengguna menyaring sejumlah besar aturan hasil penggalian untuk menemukan aturan-aturan yang penting. Permasalahan ini dapat diatasi dengan melakukan proses penggalian aturan hanya pada frequent closed itemsets. Di lain pihak, penggunaan minimum support yang sama (tunggal) untuk semua item secara implisit mengasumsikan bahwa semua item dalam basis data memiliki sifat dan frekuensi yang sama. Padahal, tidak demikian dalam kenyataannya; item yang berbeda memiliki kriteria yang berbeda pula untuk mempertimbangkan kepentingannya. Dalam mengatasi persoalan ini, diperlukan model aturan asosiasi yang memungkinkan pengguna untuk menggunakan multiple minimum support guna memperoleh gambaran mengenai sifat dasar dan frekuensi dari item-item yang ada dalam sebuah transaksi. Penelitian yang dilakukan ini berkaitan dengan pengembangan struktur Multiple Item Support Tree (MIS-tree) dan algoritma CLOSET untuk menggali frequent closed itemsets dengan menggunakan multiple minimum support. MIS-tree adalah struktur pohon yang dikembangkan serupa dengan struktur Frequent Pattern Tree (FP-tree) untuk menyimpan informasi yang ringkas (compressed) dan penting tentang frequent pattern, sedangkan algoritma CLOSET merupakan metode penggalian frequent closed itemsets berbasis proyeksi FP-tree yang efisien pada basis data besar dengan single minimum support. Algoritma yang dikembangkan diharapkan dapat menggali frequent closed itemsets dengan lebih cepat dengan menggunakan multiple minimum support. Hasil uji coba menunjukkan bahwa jumlah frequent closed itemsets yang berhasil digali berkurang ketika digunakan multiple minimum support. Hal ini menunjukkan bahwa penggunaan multiple minimum support mampu menaikkan efisiensi, yaitu berkurangnya jumlah frequent closed itemsets yang akan berpengaruh pada proses association rule mining selanjutnya. Kecepatan waktu komputasi penggalian dengan multiple minimum support juga meningkat dibandingkan dengan kecepatan penggalian menggunakan minimum support tunggal.
==================================================================================================================================
Frequent pattern mining plays an essential role in association rule mining. However, frequent pattern mining often generates a very large number of frequent itemsets and rules, which reduces the efficiency and effectiveness of the mining process since users have to sift through a vast number of mined rules to find useful ones. This problem can be addressed by mining rules only within frequent closed itemsets. On the other hand, using a single minimum support implicitly assumes that all items in the database share the same nature or similar frequencies. In real-life applications, this is often not the case; different items may have distinct criteria regarding their significance. To solve this problem, association rule analysis must be extended to allow users to specify multiple minimum supports to reflect the varying nature and frequencies of items. This research focuses on developing the Multiple Items Support Tree (MIS-Tree) structure combined with the CLOSET algorithm to mine frequent closed itemsets using multiple minimum supports. The MIS-tree is an FP-tree-like structure designed to efficiently store compressed and useful information about frequent patterns. Meanwhile, the CLOSET algorithm is an FP-tree-based database projection method used to efficiently mine frequent closed itemsets in large databases with a single minimum support. The algorithm developed in this research is particularly designed to accelerate the mining process of frequent closed itemsets using multiple minimum supports. Experimental results show that the number of mined frequent closed itemsets decreases when multiple minimum supports are applied. This in turn reduces the computing time required for subsequent association rule mining from the resulting frequent closed itemsets. This is supported by the fact that the computing time consumed by the algorithm when mining frequent closed itemsets using multiple minimum supports was faster than that required by similar algorithms using a single minimum support.

Item Type:	Thesis (Masters)
Additional Information:	RTIf 006.312 Pur p
Uncontrolled Keywords:	data mining, association rule mining, frequent closed itemsets, multiple minimum support, data mining, association rule mining, frequent closed itemsets, multiple minimum support.
Subjects:	Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science)
Divisions:	Faculty of Information Technology > Informatics Engineering > 55101-(S2) Master Thesis
Depositing User:	magang .
Date Deposited:	18 Sep 2026 08:35
Last Modified:	18 Sep 2026 08:35
URI:	http://repository.its.ac.id/id/eprint/144697

Actions (login required)

View Item