Short-term electrical load forecasting using least squares support vector machines

Abstract

Abstruct-This paper presents a Least Squares Sulpport Vector Machines (LS-SVM) approach to short-term electric load forecasting (STLF). The proposed algorithm is more robust and reliable as compared to the traditional approach when actual loads are forecasted and used as input vanabbes. In order to provide the forecasted load, the LS-SVM interpolates among the load and temperature data in a training dat;a set. Analysis of the experimental results proved that this approach can achieve greater forecasting accuracy than the tradilional model. Index Terms-Power system; Load forecasting; ]Least squares support vector machines. 0-7803-7459-2/02/$17.00 0 2002 IEEE

Electric Load Forecasting Using Support Vector Machines Optimized by Genetic Algorithm

“…Fi tn ess = MA PE(Lactual Lforecasted) (7) The genetic algorithm initialized the value of C and c then trained the SVM on training data, then GA calculated the MAPE of validation data (i.e. Fitness), it tries to minimize the MAPE by generating the new population using crossover and mutation operators.…”

“…SVM are trained with quadratic programming which gives the globally optimal solution. Least squares support vector machines for short-term load forecasting was used by [7] for 24 hour ahead load forecasting. Hai-Shan et al used the least squares support vector machines and chaos theory [8] for load forecasting and found in comparison a very promising results.…”

Bed temperature modeling and forecasting of Circulating Fluidized Bed boiler based on SMO algorithm

“…Recently, a new machine learning methods -support vector machine regression(SVR)method were being widespread concerned [3]. Support vector machine is a machine learning method proposed by Vapnik and others according to statistical theory, which is based on VC dimension and structural risk minimization principle in statistical learning theory, it has a better ability to deal with the practical problems with the small samples, non-linear , high dimension and local minimum points and so on.…”

“…The sequential minimal optimization (SMO) algorithm is proposed by Platt [3][4], the main idea is dived a large optimization problem into a series small optimization problem which containing only two variables. At first, the algorithm is mainly used for classification problems, and later Smola and Schölkopf [6] proposed a training method for SVM regression which was called SMO algorithm, this is the analogy extend of the Platt algorithm.…”

Analysis of Partial Discharge Measurement Data Using a Support Vector Machine

“…In most applications, SVM has been shown to outperform other learning algorithms, especially in finding the discriminating function. Due to its promising performance, it has also been applied into problems encountered in power system industry such as in load forecasting [14], power system stability [15][16], and fault location detection [17]. SVM algorithm generates function from a set of labeled training data.…”