基于抗差估计的概率积分法的预计参数模型研究

Study of Robust Estimation Parameters Model for Probability Integral Method

针对利用实测数据求取概率积分法预计参数时,受异值或粗差影响使参数严重偏离真值,预测结果失真的问题,建立了抗差估计的概率积分法求参模型。利用谢一矿矿区开采沉陷实测资料进行人工异值干扰求参实验,结果表明,该方法抑制了异值或粗差对求参结果的影响,保证了求参结果的真实性和可靠性。

In order to solve the divergence problem in normal least - squares fitting method because of blunder in observation data, the author develop the Robust estimation Parameters Model. A test in artificial blunder disturbance was made on the actual mining subsidence observation data in WangFengGang mining. The test resuits show that the technique of Robust estimation can reduce the disturbance of blunders, and ensure the prediction parameters authenticity and reliability.