A combined method to estimate wind speed distribution based on integrating the support vector machine with firefly algorithm

Abstract

A new hybrid approach by integrating the support vector machine (SVM) with firefly algorithm (FFA) is proposed to estimate shape (k) and scale (c) parameters of the Weibull distribution function according to previously established analytical methods. The extracted data of two widely successful methods utilized to compute parameters k and c were used as learning and testing information for the SVM-FFA method. The simulations were performed on both daily and monthly scales to draw further conclusions. The performance of SVM-FFA method was compared against other existing techniques to demonstrate its efficiency and viability. The results conclusively indicate that SVM-FFA method provides further precision in the predictions. Nevertheless, for daily estimations, the applicability of proposed method could not be feasible owing to high day-by-day fluctuations of parameters k, whereas the results of monthly estimation are completely appealing and precise. In summary, the SVM-FFA is a highly viable and efficient technique to estimate wind speed distribution on monthly scale. It is expected that the proposed method would be profitable for wind researchers and experts to be used in many practical applications, such as evaluating the wind energy potential and making a proper decision to nominate the optimal wind turbines. V C 2015 American Institute of Chemical Engineers Environ Prog, 35: 867-875, 2016

Patient-Adapted and Inter-Patient Ecg Classification Using Neural Network and Gradient Boosting

“…If this random number is smaller than p i , then generate a new nectar source according to Eq. (12). In the searching process, the nectar source X i can not find a better new nectar source after several iterations, then X i is abandoned, the corresponding employed bees change to scout bees.…”

Fuzzy Multi-Objective Optimization Algorithms for Solving Multi-Mode Automated Guided Vehicles by Considering Machine Break Time and Artificial Neural Network

“…These prediction methods usually use historical data, through some linear models include autoregressive moving average model (ARMA) [9,34], autoregressive integrated moving average model (ARIMA) [2]. The nonlinear model include SVM [8,12], LSSVM [36,39], artificial neural network (Elman neural network [44,45], echo state network [38], fuzzy neural network [6,30], RBF neural network [4,23], and etc to predict short-term wind speed. The results of some related literatures indicate that the short-term wind speed has strong nonlinearity [1,24], so the nonlinear model is more suitable for shortterm wind speed prediction.…”

“…The proposed prediction method has better regression prediction ability for short-term wind speed time series. In order to further compare the predictive effects, the proposed prediction method is compared with ARIMA [2], SVM [12], LSSVM [36], and RBF neural network [23], the simulation results are shown in Fig. 4.…”

The strategy of investment in the stock market using modified support vector regression model

“…Ahlburg [28] indicates that the MAPE is helpful in comparing various forecasting models. The MAPE has been used widely for measuring forecasting accuracy [8,[29][30][31]; thus, the MAPE is used as the first alternative accuracy measurement to evaluate the performance of the forecasting methods in this paper. The MAPE is defined in Eq.…”

“…A lower MAPE value indicates a more accurate forecasting power. On the other hand, the RMSE, which is also widely used to verify predictive accuracy [31,32], is the second performance measure to test the forecasting accuracy of the MSVR model and that of the existing models. The equation is shown in the following:…”