人工冻土温度场的支持向量机法预测

 在冻结法凿井施工中,及时掌握人工冻土温度场发展，对于优化冻结方案、处理紧急情况有重要的意义。针对新建矿井实测样本较少的特点，利用基于结构风险最小化原理与小样本学习方法-支持向量机算法，建立了人工冻土温度场发展的支持向量机计算模型。采用不同的核函数的支持向量机对人工冻土温度场发展进行对比分析，确定了适合于人工冻土温度场发展计算的核函数。人工冻土温度场支持向量机模型计算结果表明，该方法是一种有效的方法，为人工冻土温度场的计算提供了一条新途径。

Study on Support Vector Machine Method Prediction of Artificial Ground Freezing Temperature Field

During a mine shaft sinking construction with a ground freezing method,to timely understand the development of the artificial ground freezing temperature field would have important significances to optimize the freezing plan and to handle the emergent conditions.According to the features of less measured samples in a new mine construction,base on the minimized principle of the structure risk and the small sample leaning method-support vector machine algorithm,a support vector machine calculation model of the artificial freezing ground temperature field development was established.The support vector machine with different kernel functions was applied to compare and analyze the development of the artificial freezing ground temperature field.The kernel function suitable for the calculation of the artificial freezing ground temperature field development was set up.The calculation results of the model showed that the method would be a effective method and could provide a new access to the calculation of the artificial freezing ground temperature field.