Conclusion: Cluster regression analysis is more effective to deal with non-normal distribution data and multi-collinearity.

• 结论：聚类回归能较好地处理非正态性资料和多重共线性。