摘要
主要利用逐步回归和Lasso 2种方法来讨论郑州白糖期货市场白糖1011远期合约价格模型的问题,对白糖的收盘价格进行建模,在挑选出显著性变量后,根据2种方法确定的模型来分别预测白糖的价格,通过比较预测结果,发现Lasso方法的预测效果明显强于逐步回归方法,从而得到更稳定、更准确的预测模型.
The liner model for price of 1011 sugar futures,traded at Zhengzhou Futures market,with other sugar futures was analyzed by stepwise regression and Lasso methods to select significant variables.Real data were analyzed and compared,which showed that the Lasso method out-performed in prediction,resulting in a more accurate and reliable model for prediction.
出处
《北京师范大学学报(自然科学版)》
CAS
CSCD
北大核心
2011年第6期551-557,共7页
Journal of Beijing Normal University(Natural Science)
基金
教育部科学技术研究重大资助项目(309007)
中央高校基本科研业务费专项资金资助项目
关键词
逐步回归
Lasso回归
郑州白糖期货
预测
期货合约
stepwise regression
Lasso regression
Zhengzhou sugar future
prediction
futures contracts