Improved Spread-Location Visualization
针对Cleveland提出的散布-位置图，指出绝对残差或平方根绝对残差不总是适用，建议先用箱线图检查残差分布并选择合适变换，使分布对称后再作图。
Cleveland (1979Cleveland ( , 1993) ) introduced the spread-location plot as a diagnostic plot suitable for many types of fitted statistical models.The spread-location plot which plots the absolute residual or square-root absolute residual vs. fitted value along with a robust loess smooth is a useful replacement for the customary practice of plotting residuals vs. fitted values.In this note, we show that neither absolute residual or square-root absolute residual is always appropriate for error distributions likely to be encountered in actual applications.Instead we recommend the simple practice of examining the distribution of the absolute residuals with a boxplot and choosing whatever transformation is necessary to obtain symmetry before constructing the spread-location plot.We conclude with an illustrative example.