加权灰色与加权统计模型在边坡变形预测中的应用研究

 利用实测变形数据建立监测模型对边坡变形规律进行预测是工程安全监测中的重要手段。对于长期发展变化的变形来说，新、旧信息反映的规律通常会有所不同，充分合理利用新的实测信息是实时掌握边坡状态、预测下一阶段趋势的必要工作。以加权的方法增强新的监测信息的作用，适当削弱旧信息的作用，比较选取权形式，对相应的变形加权回归模型和加权灰色模型进行建模分析，其预测效果比不加权模型有较明显改善。并通过权重的改变对模型的拟合、预测结果进行了比较，总结其规律，提出了使用建议。

Applications of Weighted Regression Model and Weighted GM in Slope Deformation Forecast

Using deformation survey data to establish monitoring models is a important way to forecast deformation regulation in slope safety monitoring. To the deformation that has a long period development, there are different regulations between new and old information. Its necessary to use new information correctly to understand slope condition in time and to forecast deformation in the next period. This paper strengthens the importance of new information by giving weight. After selecting the weight function, the weighted regression models and weighted Grey Models of slope deformation have been established. Their forecast results are better than those obtianed by non-weighted models. In addition, several models have been established considering different weights. The corresponding rules and advices have been presented.