基于支持向量机的矿井排水量预测

Mine Water Discharge Forecast Based on Support Vector Machine

支持向量机是基于结构风险最小化原理的一种学习技术,具有很好泛化能力的预测工具,它有效地解决小样本、非线性、高维数、局部极小等问题。矿井排水量受降雨、河流、含水层等自然因素和煤矿开拓面积的扩大、水平的延伸等人为因素的影响,同时矿井水年排水量是非线性的时间序列。利用支持向量回归机对矿井排水量进行预测,并通过实验与文献[1]利用神经网络预测的结果进行比较,表明支持向量机具有更高的预测精度。

Support Vector Machine is a learning technology based on structure risk minimization and a predictive tool with better generalization ability, and it effectively solute the fewer samples, nonlinear, high dimension and local minima. Mine water discharge not only is influenced by rainfall, river and water strata, but also is influenced by mine exploitation areas expanding. The annual mine discharge time series is non-linear. Discharge using support vector machine is predicted, And it compare the results using neural networks through experiment, Experiment sho.w that SVM higher forecast accuracy.