尾矿坝浸润线数据挖掘预测模型的样本选取研究

本文分别应用时间序列功能模型和回归模型，在原始数据的基础上建立样本，并运用支持向量回归机算法对样本进行训练，得出了尾矿坝浸润线埋深预测模型并进行了实例应用。研究证明，运用时间序列模型选取训练样本能够得出更为精确的预测结果。

Research of selecting the training samples for the infiltration route prediction model in tailing

In this paper, based on the construction of a reasonable target system, training samples were set up by two different methods, one is multi-factor time series, the other is the traditional method. Then support vector regression algorithm machine was applied to set up two different regression models for the prediction of the future infiltration route data in tailing. The research showed that, more precise data can be obtained for the infiltration route in tailing by the time series prediction model.