Statistical Downscaling Using Kernel Quantile Regression to Predict Extreme Rainfall

Authors

  • Annisa Eki Mulyati Department of Statistics, Faculty of Mathematics and Natural Science, Bogor Agriculture University, Bogor, West Java, Indonesia
  • Aji Hamim Wigena Department of Statistics, Faculty of Mathematics and Natural Science, Bogor Agriculture University, Bogor, West Java, Indonesia
  • Anik Djuraidah Department of Statistics, Faculty of Mathematics and Natural Science, Bogor Agriculture University, Bogor, West Java, Indonesia

Keywords:

curse of dimensionality, kernel trick, quantile regression, rainfall, statistical downscaling

Abstract

Rainfall is one of climate elements with diverse intensity. In extreme circumtances it is necessary to study extreme rainfall to minimize impacts that may occur. Statistical downscaling is a method that can be used to predict rainfall, which is utilizing Global Circurlation Model (GCM) output data. The characteristics of GCM output data is curse of dimensionality which causes multicollinearity. Kernel trick is one method that can be used to overcome this problem by transforming GCM output data into a high-dimensional feature space. The transformation results are modeled with kernel quantile regression. This paper presents the use of kernel quantile regression to predict extreme rainfall, compared to kernel quantile regression with principal components. The result showed that based on the RMSEP values and the correlations, both models gave relatively similiar prediction.

References

Wigena AH. 2006. “Pemodelan statistical downscaling dengan regresi projection pursuit untuk peramalan curah hujan bulanan kasus curah hujan di Indramayu” [disertasi]. Bogor (ID): Institut Pertanian Bogor.

Djuaridah A, Wigena AH. 2011. “Regresi Kuantil untuk Eksplorasi Pola Curah Hujan di Kabupaten Indramayu”. Jurnal Ilmu Dasar 12(1): 50-56

Mondiana YQ. 2012. “Pemodelan statistical downscaling dengan regresi kuantil untuk pendugaan curah hujan ekstrim” [Tesis]. Bogor (ID): Institut Pertanian Bogor.

Cahyani TBN, Wigena AH, Djuraidah A. 2016. “Quantile regression with elastic net in statistical downscaling to predict extreme rainfall”. Glob J Pure Appl Math. 12(4): 3517–3524.

Zaikarina H, Djuraidah A, Wigena AH. 2016. “Lasso and ridge quantile regression using cross validation to estimate extreme rainfall”. Glob J Pure Appl Math. 12(3): 3305–3314.

Goldameir NE, Djuraidah A, Wigena AH. 2015. “Quantile spline regression on statistical downscaling model to predict extreme rainfall in Indramayu”. Hikari J Appl Math Sci. 9(126):6263-6272.

Khairunisa, Y. 2015. “Pemodelan Support Vector Machine Quantile Regression Untuk Prediksi Curah Hujan Bulanan Pada Musim Kemarau Studi Kasus Kabupaten Indramayu” [Thesis]. Bogor (ID): Institut Pertanian Bogor.

Bousquet, Perez-Cruz. 2004. “Kernel Methods and Their Potential Use in Signal Processing”. IEEE Signal Processing Magazine. 21(3); 57-65

Takeuchi I, Le QV, Sears T, Smola AJ. 2006. “Nonparametric Quantile Estimation”, Journal of Machine Learning Research 7: 1231-1264

Bishop, C. M. 2006. “Pattern Recognition and Machine Learning”. New York: Springer

Buhai S. 2005. “Quantile Regression Overview and Selected Application”. Ad Astra 4. Tersedia pada: www.ad-astra.ro/journal.

Wigena AH, Djuraidah A, Sahriman S. 2015. “Statistical downscaling dengan pergeseran waktu berdasarkan korelasi silang”. JMG. 16(1):19-24.

Karatzoglou A, Smola A, Hornik K, Zeileis A. 2004. “Kernlab-an S4 package for kernel methods in R. J Stat Softw”. 11(9): 1-20.

Gagliardini P, Scaillet O. 2012. “Tikhonov regularization for nonparametric instrumental variable estimators”. J Econometics . 167(1): 61-75.

Downloads

Published

2018-10-21

How to Cite

Mulyati, A. E., Wigena, A. H., & Djuraidah, A. (2018). Statistical Downscaling Using Kernel Quantile Regression to Predict Extreme Rainfall. International Journal of Sciences: Basic and Applied Research (IJSBAR), 42(2), 1–9. Retrieved from https://gssrr.org/index.php/JournalOfBasicAndApplied/article/view/9419

Issue

Section

Articles