Abstract: Power industry is the basic energy sector of national economy and plays an important support role for development of other industries. This paper studies the forecasting method of the output value of industrial enterprises above designated size in terms of power consumption. Based on the data of power consumption and output value of nearly 40 000 industrial enterprises above designated size, the forecasting model is built by the support vector machine optimized by particle swarm optimization (PSO-SVM). By using the data from January 2010 to July 2013 as the training sample of SVM, a forecasting and testing is made on the data from August 2013 to December 2013 and a comparison is also conducted between conventional SVM model and BP Neural Networks model. Result of the simulation shows that the PSO-SVM model can be more accurate and reliable than both SVM model and BP model in forecasting the industry output, and the forecasting method of industry output based on power consumption is scientific and feasible.

Key words: power, power supply, power consumption, GDP, SVM, PSO, forecast of output value, parameters optimization