高斯过程机器学习方法在地下水位预测中的应用

Application of Gaussian Process Machine Learning to the Prediction of Groundwater Level

高斯过程是近年来发展起来的一种新的机器学习方法,它有着严格的统计学习理论基础,对处理高维数非线性小样本复杂问题具有良好的适应性。针对地下水位的动态变化是一个复杂的非线性过程,地下水位与其影响因素之间存在着复杂的非线性关系,将高斯过程机器学习方法应用于地下水位预测,并提出相应的模型。工程实例研究表明,地下水位预测的高斯过程机器学习模型是科学可行的,预测精度高,简单实用,对地下水位预测的问题具有较好的适用性。

Gussian process （GP） is a newly developed machine learning method based on the strict statistical learning theory. GP is capable of solving the highly nonlinear problem with small samples and high dimensions. Considering that it is still difficult to reasonably determine the groundwater level because the process of dynamic groundwater level fluctuating is a complicated non-linear process, the model based on GP is proposed for predicting groundwater level by establishing a non-linear relationship between groundwater level and its impact factors. The results of a case study show that GP model is feasible, and can forecast groundwater precisely and simply. GP machine learning is fit to predict the groundwater level.