基于Copula的回归预测

REGRESSION PREDICTION BASED ON COPULA

针对回归预测问题,分别引入Copula回归函数和Copula τ分位数来对因变量进行点预测和区间预测,相应的通过均方误差和区间的平均长度作为预测准确性的评价指标,最后通过实证研究并与线性回归预测做比较表明：基于Copula的回归预测方法效果更好.

Aiming at regression prediction problems,we introduce Copula regression function and Copula τ quantile respectively,which is applied to the point prediction and the interval prediction for the dependent variable.Mean square error and the average length of interval are used as evaluation index of the prediction accuracy accordingly.At last,through empirical research and the linear regression prediction for contrast,it shows better effect for the method of regression prediction based on Copula.