Parameters optimization of air conditioning load prediction model based on PSO-SVR

Due to the selection of appropriate model parameters on SVR is difficult and it has significant effects on the performance of air conditioning load forecasting, particle swarm optimization algorithm is proposed, which is used to optimize the model parameters, replacing the traditional traversal method and genetic algorithm. The study result was verified by the Trail-2 benchmark data provided by the Society of Sanitary Engineers, who held an open benchmark test on the heat load prediction in 1997, PSO takes very less time as Compared with the traversal method and the predicted result satisfies the level of measurement requirement as EEP (Expected Error Percentage) was adopted as the evaluation of the prediction accuracy. the results showed that the new method not only can assure the prediction precision but also can reduce training time markedly.