基于岭回归和Lasso回归的螺纹钢期货价格实证分析

Empirical analysis of rebar futures price based on ridge regression and Lasso regression

期货是金融市场的重要组成部分,期货的交割通常是在一段时间后进行,因此对于期货价格预测显得尤为重要,其中螺纹钢期货价格预测成为提高我国钢铁产业竞争力的重要举措.通过对螺纹钢期货结算价及其影响因素的数据进行分析,分别利用岭回归和Lasso回归两种方法消除共线性的影响,得到两种修正多元回归模型,应用两种修正多元回归模型分别对螺纹钢未来一周的期货价格进行预测,并与真实价格进行比较,最终发现基于两种方法得到的模型预测准确率均高于95%以上,且基于Lasso回归方法的拟合效果更好,证明构建的两种回归模型对螺纹钢价格的走势与预测均有重要的参考价值.

Futures were an important part of the financial market.The delivery of futures was usually carried out after a period of time,so it was particularly important for futures price forecasting.The rebar forecasting of rebar futures was an important measure to improve the competitiveness of China s steel industry.The data of the rebar futures settlement price and its influencing factors were analyzed by using the two methods of ridge regression and Lasso regression to eliminate the influence of collinearity.Two modified multiple regression models were applied respectively.The futures price of rebar in the future was predicted and compared with the real price.Finally,the prediction accuracy of the model based on the two methods was higher than 95%,and the fitting effect based on the Lasso regression method was better.It was proved that the two regression models constructed have important reference value for the trend and prediction of rebar price.