基于邻域粗糙集和支持向量机的固结系数预测

采用邻域粗糙集和支持向量机建立滹沱河某地区软土固结系数的预测模型。基于自行改装的渗透固结仪,利用公式法确定不同压力下的固结系数。通过室内试验确定土体的指标参数,采用邻域粗糙集对该指标参数进行属性约简,将约简后的指标参数作为影响因素,分别建立支持向量机和神经网络的固结系数预测模型,预测未知样本的固结系数,并与实测值进行对比。结果表明:公式法可以准确客观地确定固结系数;支持向量机和BP神经网络建立的该地区软土固结系数预测模型均可以预测区域内未知点的固结系数,且支持向量机方法的预测精度比神经网络方法的预测精度提高了约10%。本文提出的方法直接从实验数据出发,通过易获取的影响因素建立特定地区固结系数预测模型,并可预测该区域其余未知点的固结系数。

Prediction of Consolidation Coefficient Based on Neighborhood Rough Set and Support Vector Machine

The prediction model of soft soil consolidation coefficient in Hutuo River was established by neighborhood rough set (NRS) and support vector machine(SVM). Based on the modified osmotic oedometer, the consolidation coefficient was determined under different pressures. The soil indices were determined by laboratory tests,and the attributes were reduced by NRS method. Subsequently, by taking the reduced indices as the influencing factors; the prediction model of consolidation coefficient by SVM and back propagation neural network (BPNN) was used to predict the consolidation coefficient of unknown samples,and the results were compared with the measured values. The results show that:the consolidation coefficient can be determined by formula method. The consolidation coefficient prediction model established by SVM and BPNN can be applied to predict the consolidation coefficients, and the prediction accuracy of SVM is about 10% higher than that of BPNN. The results are validated by the standard solution. In summary, the method presented is based on experimental data, the prediction model of consolidation coefficient is established through easily accessible factors, and can predict the consolidation coefficient of other unknown points in this area.