摘要
通过选取黄金价格、原油价格、美元指数、美国非农失业率、路透CRB指数以及白银ETF持仓量等6个变量,运用回归原理对白银价格走势进行预测.通过回归诊断发现,自变量之间存在着比较严重的多重共线性,为消除自变量共线性产生的不利影响,运用岭回归模型对白银价格走势进行预测,拟合度较好.
Six variables such as the price of gold and crude oil, the dollar index, US non-agricultural unemployment rate, Reuters CRB index and ETF were selected to predict the silver price trend by using regression theory. Through the regression diagnosis, serious multicollinearity was found between variables. In order to eliminate the bad influence on variables muhicollinearity, the ridge regression modelwas established topredict the silver price trend, and the model fit better.
出处
《高师理科学刊》
2015年第5期11-14,共4页
Journal of Science of Teachers＇College and University
关键词
白银价格
多重共线性
岭回归
silver price
muhicollinearity
ridge regression
