Open AccessPre-processing Streamflow Data through Singular Spectrum Analysis with Fuzzy C-Means Clustering
Author(s) -
Najah Nasir,
Ruhaidah Samsudin,
Ani Shabri
Publication year - 2020
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/864/1/012085
Subject(s) - autoregressive integrated moving average , streamflow , mean squared error , moving average , fuzzy logic , singular spectrum analysis , cluster analysis , data mining , statistics , mathematics , computer science , time series , geography , artificial intelligence , cartography , drainage basin , singular value decomposition
One approach to improve water resource management is by making use of streamflow forecasts. In this study, eigenvector pairs were clustered by employing fuzzy c-means (FCM) during the grouping stage as an enhancement to the singular spectrum analysis (SSA) technique for data pre-processing. The FCM-SSA pre-processed streamflow data was then supplied to an auto-regressive integrated moving average (ARIMA) model for forecasting. The Department of Irrigation and Drainage Malaysia provided the monthly streamflow records of Sungai Muda (Jambatan Syed Omar) and Sungai Muda (Jeniang) for this research, wherein each was split into training (90%) and testing (10%) sets. The R software was the platform for building every FCM-SSA-ARIMA, SSA-ARIMA and ARIMA model, while the root mean squared errors and mean absolute errors were used to compare the performance between those models. The proposed FCM-SSA-ARIMA was discovered to be capable of surpassing the SSA-ARIMA and ARIMA models.