说明：  高斯过程回归是一种基于贝叶斯原理的统计机器学习方法，将先验分布通过贝叶斯定理转化成后验分布，与其他没有采用贝叶斯技巧的预测方法而言，高斯过程最大的优点是能方便地推断出超参数，同时也能方便地给出预测值的置信区间
(Gaussian Process Regression is a statistical machine learning method based on Bayesian principle. It transforms prior distribution into posterior distribution by Bayesian theorem. The greatest advantage of Gauss process regression is that it can easily infer Super-parameters and give confidence interval of predicted values.)