Prediction of Weekly Rainfall in Semarang City Use Support Vector Regression (SVR) with Quadratic Loss Function

Alan Prahutama, Hasbi Yasin



Semarang city is one of the busiest city in Indonesia. Doe to its role as the capital city of Central Java, Semarang is known as having a relativity high rate economic activities. The geographic of Semarang city bordered by the Java sea, thus whenever the rainfall is high, there could be flood at certain area. Therefore, prediction of rainfall is very important. Support vector machine (SVM) is one of the most popular methods in nonlinear approach. One of the branches of this method for prediction is support vector regression (SVR). SVR can be approached by quadratic loss function. The study is focus on Semarang rainfall prediction during 2009 to 2013 using several kernel function. Kernel Function can provide optimal weight Some of kernel functions are linear, polynomial, and Radial Basis Function (RBF). Using this method, the study provide 71.61% R-square in the training data, for C parameter 2 with polynomial (p=2), and 71.46% R-square for the testing data



Rainfall; support vector regression (SVR); kernel function

Full Text:



  • There are currently no refbacks.

Published by Department of Chemical Engineering University of Diponegoro Semarang
Google Scholar

IJSE  by is licensed under Creative Commons Attribution 3.0 License.