NO.PZ2020010801000027
问题如下：
Define homoscedasticity and heteroskedasticity. When might you expect data to be homoscedastic?
选项：
解释：
Homoscedasticity is a property of the model errors where they have the same variance. Heteroskedasticity is a property of the errors where their variance changes systematically with the explanatory variables in the model. Experimental data are highly likely to be homoscedastic. In general, the more homogeneous the data, the more likely the errors will be homoscedastic. In finance, we often use data with substantially different scales, for example, corporate earnings or leverage ratios. This heterogeneity is frequently accompanied by heteroskedas-ticity in model errors.
为什么规模相差很大的数据会导致异方差性？如果对做一些处理，例如取log会防止异方差性吗？取log会带来什么后果？
