Study Of The Times Series Model In Aluminum Price Prediction

Aluminum industry is a pillar industry of the national economy,featuring an important strategic position.As an important basis energy industry,the economic trend of the aluminum industry is directly responsed by market conditions.However,the relationship between the variables in the market is very complicated,an effective way to grasp the overall market is price analysis.The price always plays a fundamental regulatory role in the development process of the entire aluminum industry,so forecasting the trend of the aluminum price is particularly important.Firstly,according to the average months price of aluminum in Yangtze River spot market from January 2006 to December 2011,this article used Time series method to analyze the aluminum market prices and this method is frequently used in the international arena.And then this article has set up short-term prediction models of aluminum price,and proved that ARMA(3,0)model is accurate to short-term forecast of aluminum through empirical test from January 2012 to March 2012.Secondly,because of that the error is larger when use ARMA(3,0)model to forecast in the long term and ignore the lack of analysis of influencing factors on aluminum,this article established the gray prediction model GM(1,1)using data from Yangtze River spot market.Finally,according to the combination of model theory,this article had given different weights to the above two models based on the standard of prediction error quadratic sum minimum through their different advantages and set up Combination forecast model.And then,this article proved the science and accuracy of the combination forecasting model through the contrast with the results of a single model.