Wijaya, MilaFebbri (2013) Pencarian Pola Perilaku Pengguna Internet Menggunakan Metode Association Rule Mining (Studi Kasus Fakultas MIPA Universitas Brawijaya). Sarjana thesis, Universitas Brawijaya.
Abstract
Universitas Brawijaya merupakan salah satu perguruan tinggi di Indonesia yang memberikan fasilitas bebas internet selama 24 jam setiap harinya. Lalu lintas penggunaan internet dapat dilihat dari file catatan yaituaccess logs. Data access logs merupakan data yang berskala besar sehingga dibutuhkan teknik khusus yang disebut penggalian data. Dalam penelitian ini, penggalian dilakukan terhadap data accesslogs fakultas MIPA Universitas Brawijaya. Penggalian data dilakukan untuk mendapatkan pola asosiasi pengguna internet berdasarkan periode semester dengan waktu transaksi yang bervariasi, yaitu awal semester, tengah semester dan akhir semester. Sistem akan melakukan pencarian aturan asosiasi dengan tahap awal adalah mencari frequent itemset dengan menerapkan algoritma Apriori. Kemudian dari frequent itemset tersebut akan dilakukan pencarian pola asosiasi dengan menggunakan teknik association rule. Pada tahap akhir, sistem akan menghitung kekuatan pola asosiasi yang terbentuk dengan menggunakan lift ratio.Berdasarkan hasil uji cobadidapatkanaturan asosiasi yang memiliki nilai support tertinggi diperoleh pada jurusan Kimia periode awal semester yaitu jika mengakses google.co.idmaka para pengguna internet memiliki kecenderungan sebesar 84.84% untuk juga mengakses google.com, dengan didukung oleh peluang kemunculannya sebesar 57.51%. Sedangkan nilai rata-rata lift ratio dari rule yang dihasilkan dengan minimum confidence 40% sebesar 1.11. Nilai lift ratio tertinggi sebesar 1.46 pada jurusan Fisika periode akhir semester dan nilai lift ratio terendah sebesar 0.89 pada jurusan Matematika periode akhir semester.
English Abstract
Brawijaya University is one of the universities in Indonesia that provides free internet facility for 24 hours. The Traffic of internet usage could be seen on access logs files. Access log data is a large-scaled data which need a special technique was called data mining. In this research, data mining were carried out on the access log data of Faculty Mathematics and Science, Brawijaya University. The data mining were performed to obtained the association pattern of internet users in a semester with different transaction time, within a beginning of the semester, mid of semester and end of semester. The system would performed a searching of association rules with the early step was find the frequent item sets by applying the apriori algorithm. Then, based on the frequent item sets, the association patterns search would be performed by using association rule technique. At the final stage, the system would calculated the strength of association patterns which were formed by using lift ratio. Based on the test results, obtained association rules which has the highest support values is in the early period of the semester in Chemistry Department, that if users accessed google.co.id, they also have a tendency of 84.84% to access google.com, which is supported by a chance occurrence of 57.51%. And 1.11 in average value of lift ratio from the rules generated, with the minimum confidence 40%. The higest value of lift ratio was 1.46 in Physics Department at the end of semester period and the lowest values was 0.89 in Mathematics Department at the end of semester period.
| Item Type: | Thesis (Sarjana) |
|---|---|
| Identification Number: | SKR/FTIK/2013/44/051301075 |
| Subjects: | 000 Computer science, information and general works > 005 Computer programming, programs, data |
| Divisions: | Fakultas Ilmu Komputer > Teknik Informatika |
| Depositing User: | Hasbi |
| Date Deposited: | 28 Mar 2013 10:01 |
| Last Modified: | 23 Oct 2021 09:48 |
| URI: | http://repository.ub.ac.id/id/eprint/145830 |
Preview |
Text
Jurnal_-_Mila_Febbri_Wijaya.pdf Download (2MB) | Preview |
Preview |
Text
Laporan_-_Mila_Febbri_Wijaya.pdf Download (6MB) | Preview |
Actions (login required)
 |
View Item |