Density Weighted Least Squares Support Vector Machine

Least squares support vector machine is works with a sum squares errors cost function which used to minimization its empirical risk. The higher distribution samples are well fitted by the model because the estimation of the support values is optimal in the case of a Gaussian distribution, but the peak samples are poor fitted for its sparse distribution. A density weighted least squares support vector machine is proposed here, which based on the weighted least squares method. In this model, the errors of sparsely distributing samples are higher weighted in the optimization function, which help to improve the fitting accuracy of peak samples significantly with the average accuracy maintained simultaneously. The feasibility and the efficacy of this model are demonstrated on function fitting and load forecast of power system in the last.