Wavelet Support Vector Machine Model Research Articles

Predictive uncertainty assessment in flood forecasting using quantile regression

Abstract Floods and their associated impacts are topics of concern in land development planning and management, which call for efficient flood forecasting and warning systems. The performance of flood warning systems is affected by uncertainty in water level forecasts, which is due to their inability to measure or calculate a modeled value accurately. Predictive uncertainty is an emerging type of uncertainty modeling technique that emphasizes total uncertainty quantified as a probability distribution conditioned on all available knowledge. Predictive uncertainty analysis was done using quantile regression (QR) for machine learning-based flood models – Hybrid Wavelet Artificial Neural Network model (WANN) and Hybrid Wavelet Support Vector Machine model (WSVM) for different lead times. Comparing QR models of WANN and WSVM revealed that the slope, intercept, spread of forecast, and width of confidence band of the WANN model are more for each quantile indicating more uncertainty as compared to the WSVM model. In both models, with an increase in lead time, uncertainty has shown an increasing trend as well. The performance evaluation of inference obtained from QR models was evaluated using uncertainty statistics such as prediction interval coverage probability, average relative interval length (ARIL), and mean prediction interval (MPI).

Sea level prediction using climatic variables: a comparative study of SVM and hybrid wavelet SVM approaches

Climate change is expected to adversely affect the coastal ecosystem in many ways. One of the major consequences of climate change in coastal areas is sea level rise. In order to manage this problem efficiently, it is essential to obtain reasonably accurate estimates of future sea level. This study focuses essentially on the identification of climatic variables influencing sea level and sea level prediction. Correlation analysis and wavelet coherence diagrams were used for identifying the influencing variables, and support vector machine (SVM) and hybrid wavelet support vector machine (WSVM) techniques were used for sea level prediction. Sea surface temperature, sea surface salinity, and mean sea level pressure were observed to be the major local climatic variables influencing sea level. Halosteric effect is found to have a major impact on the sea level. The variables identified were subsequently used as predictors in both SVM and WSVM. WSVM employs discrete wavelet transform to decompose the variables before being input to the SVM model. The performance of both the models was compared using statistical measures such as root mean square error (RMSE), correlation coefficient (r), coefficient of determination (r2), average squared error, Nash–Sutcliffe efficiency, and percentage bias along with graphical indicators such as Taylor diagrams and regression error characteristic curves. Results indicate that the WSVM model predicted sea level with an RMSE of 0.029 m during the training and 0.040 m during the testing phases. The corresponding values for SVM are 0.043 m and 0.069 m, respectively. Also, the other statistical measures and graphical indicators suggest that WSVM technique outperforms the SVM approach in the prediction of sea level.

Wavelet-support vector machine for forecasting palm oil price

This study examines the feasibility of applying Wavelet-Support Vector Machine (W-SVM) model in forecasting palm oil price. The conjunction method wavelet-support vector machine (W-SVM) is obtained by the integration of discrete wavelet transform (DWT) method and support vector machine (SVM). In W-SVM model, wavelet transform is used to decompose data series into two parts; approximation series and details series. This decomposed series were then used as the input to the SVM model to forecast the palm oil price. This study also utilizes the application of partial correlation-based input variable selection as the preprocessing steps in determining the best input to the model. The performance of the W-SVM model was then compared with the classical SVM model and also artificial neural network (ANN) model. The empirical result shows that the addition of wavelet technique in W-SVM model enhances the forecasting performance of classical SVM and performs better than ANN.

A Data-Driven-Based Wavelet Support Vector Approach for Passenger Flow Forecasting of the Metropolitan Hub

With the rapid development of the construction and operation of mass transit hubs, passenger data collection, modeling, and prediction for optimal control have become very important. In this paper, pedestrian facilities are abstracted into connected nodes, and the passenger flow network is formed according to the facility connection relationship determined by the traffic organization; therefore, the state variables of the hub, such as saturation, and the traveling time can be estimated by pedestrian flow information collected by camera monitors and a free Wi-Fi network, including the fast analysis of data features and traffic flow prediction. The method is applied to a real case. The features of pedestrian flows are classified as chaotic and nonchaotic. We use a regression model to predict the nonchaotic situation, and the wavelet support vector machine model is proposed for the chaotic. The results can be used for the control of exits and ramps in the hub.

River discharge simulation using variable parameter McCarthy–Muskingum and wavelet-support vector machine methods

In this study, an extended version of variable parameter McCarthy–Muskingum (VPMM) method originally proposed by Perumal and Price (J Hydrol 502:89–102, 2013) was compared with the widely used data-based model, namely support vector machine (SVM) and hybrid wavelet-support vector machine (WASVM) to simulate the hourly discharge in Neckar River wherein significant lateral flow contribution by intermediate catchment rainfall prevails during flood wave movement. The discharge data from the year 1999 to 2002 have been used in this study. The extended VPMM method has been used to simulate 9 flood events of the year 2002, and later the results were compared with SVM and WASVM models. The analysis of statistical and graphical results suggests that the extended VPMM method was able to predict the flood wave movement better than the SVM and WASVM models. A model complexity analysis was also conducted which suggests that the two parameter-based extended VPMM method has less complexity than the three parameter-based SVM and WASVM model. Further, the model selection criteria also give the highest values for VPMM in 7 out of 9 flood events. The simulation of flood events suggested that both the approaches were able to capture the underlying physics and reproduced the target value close to the observed hydrograph. However, the VPMM models are slightly more efficient and accurate, than the SVM and WASVM model which are based only on the antecedent discharge data. The study captures the current trend in the flood forecasting studies and showed the importance of both the approaches (physical and data-based modeling). The analysis of the study suggested that these approaches complement each other and can be used in accurate yet less computational intensive flood forecasting.

Wavelet and cuckoo search-support vector machine conjugation for drought forecasting using Standardized Precipitation Index (case study: Urmia Lake, Iran)

Abstract In this study, wavelet-support vector machine (WSVM) is proposed for drought forecasting using the Standardized Precipitation Index (SPI). In this way, the SPI time series of Urmia Lake watershed is decomposed to multiple frequency time series by wavelet transform. Then, these time sub-series are applied as input data to the support vector machine (SVM) model to forecast drought. Also, a cuckoo search (CS)-based approach is proposed for parameter optimization of SVM, finding the best initial constant parameters of the SVM algorithm. The obtained results indicate that the radial basis function (RBF)-kernel function of the SVM algorithm has high efficiency in the SPI modeling, resulting in a determination coefficient (DC) of 0.865 in verification step. In the WSVM model, the Coif1, which is considered as a mother wavelet function with decomposition level of five, shows a better performance with DC of 0.954 in verification step, revealing that the proposed hybrid WSVM model outperforms the single SVM model in forecasting SPI time series. Also, DC of cuckoo search-support vector machine (CS-SVM) is calculated to be 0.912 in verification step, indicating the fact that the proposed CS-SVM model shows better efficiency than single SVM model.

A Wavelet Support Vector Machine Combination Model for Singapore Tourist Arrival to Malaysia

In this study, wavelet support vector machine model (WSVM) is proposed and applied for monthly data Singapore tourist time series prediction. The WSVM model is combination between wavelet analysis and support vector machine (SVM). In this study, we have two parts, first part we compare between the kernel function and second part we compare between the developed models with single model, SVM. The result showed that kernel function linear better than RBF while WSVM outperform with single model SVM to forecast monthly Singapore tourist arrival to Malaysia.

Hybrid wavelet-support vector machine approach for modelling rainfall–runoff process

Because of the importance of water resources management, the need for accurate modeling of the rainfall-runoff process has rapidly grown in the past decades. Recently, the support vector machine (SVM) approach has been used by hydrologists for rainfall-runoff modeling and the other fields of hydrology. Similar to the other artificial intelligence models, such as artificial neural network (ANN) and adaptive neural fuzzy inference system, the SVM model is based on the autoregressive properties. In this paper, the wavelet analysis was linked to the SVM model concept for modeling the rainfall-runoff process of Aghchai and Eel River watersheds. In this way, the main time series of two variables, rainfall and runoff, were decomposed to multiple frequent time series by wavelet theory; then, these time series were imposed as input data on the SVM model in order to predict the runoff discharge one day ahead. The obtained results show that the wavelet SVM model can predict both short- and long-term runoff discharges by considering the seasonality effects. Also, the proposed hybrid model is relatively more appropriate than classical autoregressive ones such as ANN and SVM because it uses the multi-scale time series of rainfall and runoff data in the modeling process.