“…Therefore, besides the regression models for multidimensional data, how to model different kinds of noises, what loss functions to use, and how to improve the robustness of regression algorithms are still current research hotspots (Yi and Faloutsos, 2000;Liu et al, 2017a;Diakonikolas et al, 2019;Luo et al, 2015;Rousseeuw and Hubert, 2018;Zhang and Zhou, 2018;Takada et al, 2019). In addition, with the emergence of some new application fields, the applicability of traditional regression methods for practical problems has also become a promising research direction, such as food safety control (Lim et al, 2016;Cheng and Sun, 2017), smart transportation (Xu and Huang, 2015;Mei et al, 2019), medical analysis (Zhu et al, 2016;Zhen et al, 2015), model performance assessment (Hewing et al, 2019), and environmental risk assessment (Kang et al, 2019;Arbaiy et al, 2018). We refer interested readers to these books (Hastie et al, 2009;Lawrence, 2019;Bates and Watts, 1988) and review paper (Fernández-Delgado et al, 2018) for more details of traditional regression analysis.…”