【摘要】 以某在建高速公路为研究背景，通过对相关系数R在Asaoka、灰色理论、双曲线和指数曲线这4种拟合算法中预测精度进行深入分析并结合工程实际情况，提出将平均绝对误差MAD、均方误差MSE和平均绝对百分误差MAPE这3个常用的评价指标结合起来使用，以此衡量沉降预测精度的改进方案，并从适用性和预测精度角度进行论述。结果表明：改进后的评价新模式比相关系数的评价方式更加接近工程实际需求，其可显著提高满足沉降预测精确度要求的测点百分比。更多还原

【Abstract】 Taking a highway under construction as the research background, through to the correlation coefficient R in Asaoka, gray theory, the four kinds of hyperbola and exponential curve fitting method of precision evaluation of in-depth analysis and combined with the engineering actual situation, proposed to MAD mean absolute error, mean square error(MSE, and mean absolute percentage error MAPE used to combine these three commonly used evaluation index to measure the settlement prediction accuracy improvement scheme, and from two angles of applicability and prediction precision are discussed. The results show that the improved evaluation new model than the correlation coefficient of evaluation way, demand closer to the engineering practice, significantly improved the percentage points that could satisfy the requirement of settlement prediction accuracy.更多还原