Rainfall-runoff modelling using adaptive neuro-fuzzy inference system

• Main Authors: Nurul Najihah, Che Razali, Ngahzaifa, Ab. Ghani, Syifak Izhar, Hisham, Shahreen, Kasim, Widodo, Nuryono Satya, Sutikno, Tole
• Format: Article
• Language: English
• Published: Institute of Advanced Engineering and Science 2020
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://umpir.ump.edu.my/id/eprint/27154/1/21145-40184-1-PB.pdf; http://umpir.ump.edu.my/id/eprint/27154/; http://doi.org/10.11591/ijeecs.v17.i2.pp1117-1126; http://doi.org/10.11591/ijeecs.v17.i2.pp1117-1126
• Institution: Universiti Malaysia Pahang

This paper discusses the working mechanism of ANFIS, the flow of research, the implementation and evaluation of ANFIS models, and discusses the pros and cons of each option of input parameters applied, in order to solve the problem of rainfall-runoff forecasting. The rainfall-runoff modelling considers time-series data of rainfall amount (in mm) and water discharge amount (in m3/s). For model parameters, the models apply three triangle membership functions for each input. Meanwhile, the accuracy of the data is measured using the Root Mean Square Error (RMSE). Models with good performance in training have low values of RMSE. Hence, the 4-input model data is the best model to measure prediction accurately with the value of RMSE as 22.157. It is proven that ANFIS has the potential to be used for flood forecasting generally, or rainfall-runoff modelling specifically.