ANIFAH, UMI (2014) Pemilihan Threshold Pada Pemodelan Nilai Ekstrim Curah Hujan Di Sentra Produksi Padi Jawa Timur Dengan Mrlp-Baru (New Technique For Mean Residual Life Plot). Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
1310100057-Abstract.pdf - Published Version Download (348kB) | Preview |
Preview |
Text
1310100057-Undergraduate Thesis.pdf - Published Version Download (3MB) | Preview |
Preview |
Text
1310100057-Conclusion.pdf - Published Version Download (614kB) | Preview |
Abstract
Perubahan iklim ekstrim tentunya berdampak pada
beberapa sektor kehidupan. Salah satu sektor yang sangat
merasakan dampak dari fenomena iklim ekstrim adalah sektor
pertanian. Kegagalan panen yang disebabkan oleh banjir dan
kekeringan masih saja selalu terjadi. Fenomena tersebut sering terjadi pula pada lima Kabupaten yang menjadi sentra produksi padi di Jawa Timur, antara lain: Kabupaten Jember,
Banyuwangi, Bojonegoro, Lamongan dan Ngawi. Untuk
mengatasi permasalahan tersebut, maka perlu dilakukan
antisipasi dini. Pengembangan ilmu dan pengetahuan perlu
dilakukan dalam kontribusi untuk memecahkan permasalahan ini.
Terdapat suatu metode yang mengkaji fenomena ekstrim yaitu
Ekstreme Value Theory (EVT). EVT merupakan metode yang
membahas distribusi nilai ekstrim. Salah satu metode EVT adalah Peaks Over Threshold (POT). Pada POT, bahasan tentang
pemilihan threshold menjadi menarik untuk dikaji. Metode
prosentase merupakan metode untuk mendapatkan threshold
yang sudah sering digunakan. Namun, ada metode lain yang
menarik untuk dibahas adalah MRLP (Mean Residual Life Plot).
Pada penelitian ini, digunakan Likelihood-Ratio test untuk
memutuskan pilihan threshold yang terbaik. Serangkaian metode dalam penelitian ini kemudian disebut dengan MRLP-baru (new technique for MRLP). Hasil penelitian menunjukkan bahwa metode MRLP-baru lebih baik daripada metode prosentase. Namun, hasil RMSE secara keseluruhan menunjukkan bahwa kedua metode tersebut masih sama. Hal ini karena adanya kemungkinan sifat dependensi data yang mempengaruhi estimasi return level. Oleh karena itu disarankan untuk penelitian selanjutnya memperhatikan syarat data yang independen.
==============================================================================================================================
Extreme climate change would have an impact on some
sectors of life. One sector that felt the impact of extreme climatic
phenomena is the agricultural sector. Crop failure caused by
floods and droughts still always the case. This phenomenon often
occurs also in the five districts that became centers of rice
production in East Java, among others: Jember, Banyuwangi,
Bojonegoro, Lamongan and Ngawi. To overcome these problems,
it is necessary to anticipate early. Development of science and
knowledge needs to be done in a contribution to solving this
problem. There is Extreme Value Theory (EVT), a method that
examines the phenomenon of extreme. EVT is a method that
addresses the extreme value distribution. One of the EVT method
is Peaks Over Threshold (POT). In POT, a discussion of the
election threshold be interesting to study. Percentage method is a
method to obtain a threshold that has been frequently used.
However, there is another interesting method to be discussed is
MRLP (Mean Residual Life Plot). In this study, the likelihoodratio
test is used to decide which option best threshold. A series of
methods in this study are referred to MRLP-new (new technique
for MRLP). The results showed that the new method MRLPpercent
better than the method. However, the overall RMSE
results indicate that both methods are still the same. This is due to
potential data dependencies that affect the nature of the return
level estimates. It is therefore recommended for further research
attention to the requirements of independent data.
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSSt 519.536 Ani p |
| Uncontrolled Keywords: | EVT, MRLP-baru, POT, return level, threshold |
| Subjects: | Q Science > QA Mathematics Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression |
| Divisions: | Faculty of Mathematics and Science > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Mr. Tondo Indra Nyata |
| Date Deposited: | 04 Jan 2017 07:27 |
| Last Modified: | 13 Nov 2024 02:20 |
| URI: | http://repository.its.ac.id/id/eprint/1301 |
Actions (login required)
 |
View Item |