0

I have a dataset where variance is different across groups, and there exist some outliers too. I used robust linear regression lmrob() to handle outliers. However, I need a method that takes care of variance heteroscedasticity too.

Also, the application of such method in R can be very helpful.

Thanks in advance for your help. :)

SteveMcManaman
  • 143
  • Dealing with heteroscedasticity is a broad topic and there is no single black box function in R that solves it. So, could you explain more about the data. For instance, what sort of heteroscedasticity do you have? I am asking because possibly you can delay with it by using a data transformation. – Sextus Empiricus Jul 05 '22 at 11:50
  • @SextusEmpiricus I'm trying to explain a numerical target variable using two categorical variables. The target variable is zero-inflated in one group of explanatory variables, and has some outliers too. I could control for the outliers using robust regression, but the diagnostic plots show heterogeneity in variances. I hope that could clarify my question. – SteveMcManaman Jul 05 '22 at 11:55
  • You have the distributions in each combination of categorical variables, right, like black/dog, white/dot, black/cat, and white/cat? What more do you hope to get from running a regression? – Dave Jul 05 '22 at 15:24
  • You might see this post for some options for handling heteroscedasticity. There's an example there of using lmrob() with weighted observations. stats.stackexchange.com/questions/91872/alternatives-to-one-way-anova-for-heteroskedastic-data/91881#91881 – Sal Mangiafico Jul 05 '22 at 15:47

0 Answers0