地下开采引发地面沉陷的未确知聚类预测方法

对未确知聚类预测法进行优化，并将其应用于开采地面沉陷的预测研究。采用开采地面沉陷的实测数据按最大沉陷量进行分类，利用各分类影响因素的均值表示各分类中心，并确定各影响因素的未确知测度函数。由待测对象指标的综合未确知测度与各分类指标的未确知测度间的距离来确定待预测对象所属等级，给出了预测值的计算公式。经计算验证，该方法的正确率为75％。但在实际应用中，为了保证地表建筑设施等更加安全，允许预测级高判，则正确率可达100％。针对某铁矿一观测点进行预测，并与实测数据比较，结果表明，未确知聚类预测的结果是令人满意的，为开采地面沉陷的预测提供了一种新思路。

Predicting surface subsidence induced by mining based on unascertained clustering method

Forecast method of unascertained clustering to predict mining induced surface subsidence is optimized. Maximal subsidence displacement is classified by practical mining induced surface subsidence data, the centre of every classified is indicated by using averages of index of every classification, and unascertained measurement function is ascertained. Gradeof waiting forecast sample is estimated based on the distance between the unascertained measurement of waiting forecast sample＇s synthesis index and that of every classification index, the formula of calculation forecasting value is also given. The computation results show that this method＇ s correct rate is 75 %, but in order to guarantee the surface facility much safer, the grade of waiting forecast sample is permitted to estimate more high, the correct rate may reach 100% in practical application. Furthermore, the forecast model of unascertained clustering is used to make prediction for existed example of one Iron Ore Mine, and the prediction is compared with real measurements. The results show that proposed model is more valid and applicable in predicting mining induced surface subsidence, idence displacement of prediction