中文摘要：

通过对修正磨光法的仿真,发现用盈亏修正磨光法和预测修正磨光法所得到的逼近效果仍然很差,最根本的原因是根据原数据进行盈亏修正磨光调整,调整的步长较小;另外,如果仅在调整值中加以调整幅度的修改,仿真发现逼近误差仍然较大.因此,这里通过把有些原始数据值进行缺失参数化后,即通过缺失数据调整修正优化磨光法所得到的绝对误差要小得多,逼近的绝对误差波动性更小.通过实例表明本算法是一种高精度,稳定性的计算方法 .

英文摘要：

This paper founds that approximation effects are still very poor through the simulation of the modified smoothing algorithm and forecasting correction smoothing algorithm too.The most fundamental reason is that the step on adjusted value of the profit and loss modified smoothing algorithm is less On the basis of original data Another, if adjustment changes are only modified in modified-value,the simulation founds that approximation error is still bigger.So,after some original data values are missed and parameterized,namely,the absolute error of result obtained by missing data adjustment modified optimization smoothing algorithm is much smaller and volatility of approximation absolute error is less.The examples show that the algorithm is a highprecision, stability calculation method.