Cost prediction of power transmission and transformation projects based on improved SVM

A power engineering cost prediction model based on improved SVM is proposed to address the problems of slow solving speed and unsatisfactory performance of support vector machine in predicting power engineering costs. This model comprehensively considers the components of power engineering costs and normalizes the parameters. It improves the SVM model using least squares estimation and uses genetic algorithm to solve the optimal values of LSSVM parameters. The optimized GA-LSSVM model is used to predict power engineering costs. The simulation experiment results based on MATLAB simulation platform show that the predicted engineering cost by the model is close to the actual value, with normalized mean square error and average absolute percentage error of 18 330000 yuan and 3.58%, and the prediction time is only 256ms, proving that its overall performance is superior to other comparative models.