Komparasi Kajian Model Hidrologi Runtun Waktu Menggunakan Soft Computing (Studi Kasus : DAS Siak Hulu)

Abstract

There are several methods that can be used to predict discharge some future time, but the results still have relatively high error value. The use of models softcomputing a method that can be used to model a hydrological analysis such as forecasting discharge. Adaptive Neuro Fuzzy Inference System (ANFIS) is one softcomputing models that have proven reliability in conducting hydrological analysis as to forecast the flood discharge, and long reservoir inflow of seawater intrusion.Reliability ANFIS models need to be tested in the analysis of hydrological especially to predict discharge.ANFIS models were built to predict discharge using the main data is data on Tapung river left in the year 2002-2012 (except on taun 2007). Then ANFIS foresees the discharge of each of the data sources and analyze the magnitude of the error forecasting results debit ANFIS models. Then the forecast results will be compared with the results of forecasting with ANN method.Results discharge by using ANFIS models showed excellent results with test parameter value statistical correlation coefficient (R) of more than 0.75 are included in the category of correlation is very strong. To forecast results debit ANFIS models, yielding a value of correlation (R) of 0.99 while the forecast results debit ANN method produces a value of correlation (R) of 0.94903.Keywords: hydrological analysis, tidal forecasting, softcomputing, ANFIS.