<![CDATA[Flood is a natural disaster that used to be general cause and hard to predict when it will happened. So far, the cause of flood is there's process when rainfall and waterlevel is rise, so there's required some research to do a monitoring on flood alert. From that point, system is required to be able to forecast and make it easier to analyze flood alert status in a future. To forecast a future results, there is a method that based on the availability of raw data, also with statistical analysis technique called regression method. Regression method that used in this research is Support Vector Regression. This SVR method is frequently used in forecasting, but not many of them use rainfall and waterlevel data in a same time. The purpose of this research is to do flood alert forecasting in Kambing Station DAS Brantas. The results represent flood alert forecasting at December 2016, with waterlevel data resulted minimal value of 9.584849544 in error rate and rainfall data resulted minimal value of 10.52259887 in error rate. By using values of parameters = 0.09, = 0.005, = 0.2, = 0.08 and = 0.08. Both data resulted flood alert forecasting that shows Normal.]]>
Copyrights © 2018
