Identification of support vector machines for runoff modelling

Abstract

This paper describes an exploration in using SVM (Support Vector Machine) models, which were initially developed in the Machine Learning community, in flood forecasting, with the focus on the identification of a suitable model structure and its relevant parameters for rainfall runoff modelling. SVM has been applied in many fields and has a high success rate in classification tasks such as pattern recognition, OCR, etc. The applications of SVM in regression of time series are relatively new and they are more problematic in comparison with classifications. This study found that exhaustive search of an optimum model structure and its parameter space is prohibitive due to their sheer size and unknown characteristics. Some parameters are very sensitive and can increase the CPU load tremendously (and hence result in very long computation times). All these make it very difficult to efficiently identify SVM models, which has been carried out by manual operations in all study cases so far. The paper further explored the relationships among various model structures (ξ-SV or ν-SV regression), kernel functions (linear, polynomial, radial basis and sigmoid), scaling factor, model parameters (cost C, epsilon) and composition of input vectors. These relationships should be able to provide useful information for more effective model identification in the future. The unit response curve from SVM was compared with a transfer function model and it is found that a TF model outperforms SVM in short-range predictions. It is still unclear how the unit response curve could be utilised for model identification processes and future exploration in this area is needed.

Support Vector Regression for Rainfall-Runoff Modeling in Urban Drainage: A Comparison with the EPA’s Storm Water Management Model

“…Input vectors contain rainfall and flow observations. However, as highlighted by Bray and Han [13], analysis results show that SVR effectiveness is less affected by rainfall and the performance enhancement of the model with increased flow observation is more significant than the improvement due to increased rainfall observation. An exhaustive search of an optimum model structure and its parameters is very complicated.…”

“…It is based on a long and tedious trial and error iteration procedure, which is very difficult to automate and is omitted here for reasons of brevity. More details can be found in Bray and Han [13]. The deviation between the target value and the function found by the SVM is controlled by the ε parameter.…”

“…Dibike et al [12] investigated SVMs for rainfall-runoff modeling compared to ANN models. Bray and Han [13] highlighted the difficulties in identifying a suitable model structure and its relevant parameters for rainfall-runoff modeling. Tripathi et al [14] showed an SVM approach for statistical the downscaling of precipitation at a monthly time scale.…”

Assessment of the impacts of nonstationarity on watershed runoff using artificial neural networks: a case study in Ardebil, Iran

“…Because of the limited resources associated with developing and calibrating conceptual, metric, and physics models (Kokkonen and Jakeman 2001), data-driven hydrological methods have been widely adopted for forecasting runoff. Different studies demonstrated the ability of ANN for runoff simulation (Dawson and Wilby 2001;Han et al 2007a, b;Bray and Han 2004;Nayak et al 2005Nayak et al , 2007. Zarghami et al (2011) used of general circulation models (GCM) to predict the climate change and the three scenarios (A1B, A2 and B1) with the horizons 2020, 2055 and 2090.…”

Using support vector machines for long-term discharge prediction

“…However, along with the introduction of Vapnik's ε insensitive loss function, SVMs have been extended to solve nonlinear regression estimation (Gunn, 1998;Smola & Schölkopf, 2004) and time series forecasting (Thissen et al, 2003). It is useful to note that the SVM is finding its way into the water sector (Liong & Sivapragasm, 2002;Bray & Han, 2004;Asefa et al, 2004) and a combination of SVM and evolutionary algorithm called EC-SVM has also been attempted recently (Yu et al, 2004). In the latter study, a shuffled complex evolution (SCE-UA) algorithm (Duan et al, 1992(Duan et al, , 1993(Duan et al, , 1994 was used to search phase space parameters (the time delay embedding dimension) and three SVM parameters.…”