Using an ANOVA and then a post-hoc Tukey HSD test, we see significant improvements at a significance threshold of 0.05 between standard regression approaches and machine learning approaches.

Using an ANOVA and then a post-hoc Tukey HSD test, we see significant improvements at a significance threshold of 0.05 between standard regression approaches and machine learning approaches.我们先进行方差分析（Analysis of Variance，ANOVA），再开展事后图基HSD检验（post-hoc Tukey HSD test），在显著性阈值为0.05的条件下，观测到标准回归方法与机器学习方法之间存在显著性能提升。