Triwansa, Lipsia Cakra and Rekyan Regasari Mardi Putri,, ST., MT and Suprapto, , ST., MT (2014) Penentu Potensi Tsunami Di Indonesia Akibat Gempa Bumi Dengan Metode Naïve Bayesian Classifier. Sarjana thesis, Universitas Brawijaya.
Abstract
Program ini merupakan sebuah gambaran penelitian mengenai penerapan metode Naïve Bayesian Classifier (NBC) dalam topik penentu potensi tsunami di Indonesia akibat gempa bumi. Penelitian sebelumnya mengenai topik tsunami sudah pernah dilakukan dengan judul “Penentuan Potensi Tsunami akibat Gempa Bumi Bawah Laut dengan Metode Modified K-Nearest Neighbor (MKNN)” pada tahun 2013 yang memiliki akurasi sebesar 73.74%. Topik tsunami sendiri dipilih karena selama ini sistem peringatan dini (early warning) tsunami di Indonesia masih berdasarkan pengamatan pergerakan gelombang melalui satelit sehingga kurang efektif. Selain itu, meskipun ada ilmu mengenai prasyarat tsunami, namun dari data asli apabila prasyarat tersebut terpenuhi belum tentu juga menimbulkan tsunami. Selain metode Naïve Bayesian Classifier (NBC), penelitian ini juga menerapkan metode Bayesian Network (BN) untuk mengimplementasikan ilmu mengenai prasyarat tsunami. Program memberikan 2 hasil rekomendasi potensi tsunami. Jenis rekomendasi potensi tsunami sendiri ada 3 jenis, yaitu tidak terjadi tsunami, waspada tsunami, dan awas tsunami. Pengujian pada program dilakukan dengan 2 jenis pengujian, yaitu pengujian perubahan jumlah data training dan pengujian perubahan komposisi data training. Pengujian perubahan jumlah data training dilakukan untuk mendapatkan rata-rata akurasi dari metode Naive Bayesian Classifier (NBC) dan penerapan ilmu prasyarat tsunami dengan metode Bayesian Network (BN). Sedangkan perubahan komposisi data training digunakan untuk mendapatkan kestabilan akurasinya. Dari hasil pengujian perubahan jumlah data training didapatkan rata-rata akurasi dari metode Naïve Bayesian Classifier (NBC) dan penerapan ilmu prasyarat tsunami dengan metode Bayesian Network (BN) sebesar 86.67% dan 75%. Sedangkan pengujian perubahan komposisi data training memberikan rata-rata akurasi 79% dan 73.33%.
English Abstract
This program is an overview of research on the application of the methods of Naïve Bayesian Classifier (NBC) in deciding the topic of a potential tsunami in Indonesia by the earthquake. Previous research on the topic of the tsunami has been done with the title "Determination of Potential Earthquake Tsunami due to the method Underwater Modified K-Nearest Neighbor (MKNN)" in 2013 which has an accuracy of 73.74%. Topics tsunami itself was chosen because during this early warning system (early warning) tsunami in Indonesia is based on satellite observations of the movement of waves through making it less effective. In addition, although there is science prerequisites regarding the tsunami, but from the original data if the prerequisites are met is not necessarily also cause a tsunami. Naïve Bayesian Classifier addition method (NBC), this study also apply the method of Bayesian Network (BN) to implement the prerequisite knowledge about tsunamis. The program gives 2 results tsunami potential recommendations. Type the potential tsunami own recommendations there are 3 types, ie tsunami, tsunami alert, tsunami and alert. The test program carried out with 2 types of tests, namely the change in the amount of training data testing and test changes in the composition of the training data. Testing changes in the amount of training data is done to get an average accuracy of Naive Bayesian Classifier method (NBC) and the application of science prerequisites tsunami by the method of Bayesian Network (BN). While changes in the composition of the training data used to obtain the stability of its accuracy. From the test results of changing total training data obtained an average accuracy of Naïve Bayesian Classifier method (NBC) and the application of science prerequisites tsunami by the method of Bayesian Network (BN) of 86.67% and 75%. While testing changes in the composition of the training data gives an average accuracy of 79% and 73.33%.
Other obstract
-
| Item Type: | Thesis (Sarjana) |
|---|---|
| Identification Number: | 004 |
| Uncontrolled Keywords: | penentu potensi tsunami, Naïve Bayesian Classifier, ilmu prasyarat tsunami, Bayesian Network-determinants of potential tsunami, Naive Bayesian Classifier, science prerequisites tsunami, Bayesian Network, |
| Subjects: | 000 Computer science, information and general works > 005 Computer programming, programs, data |
| Divisions: | Fakultas Ilmu Komputer > Teknik Informatika |
| Depositing User: | Hasbi |
| Date Deposited: | 07 Aug 2014 14:16 |
| Last Modified: | 07 Dec 2021 01:22 |
| URI: | http://repository.ub.ac.id/id/eprint/145946 |
Preview |
Text
Lipsia_Cakra_T._105060801111043_Laporan_Skripsi.pdf Download (20MB) | Preview |
Actions (login required)
 |
View Item |