From the course: Machine Learning & AI Foundations: Linear Regression

Unlock the full course today

Join today to access over 22,400 courses taught by industry experts or purchase this course individually.

Dealing with outliers: Should cases be removed?

Dealing with outliers: Should cases be removed? - SPSS Tutorial

From the course: Machine Learning & AI Foundations: Linear Regression

Start my 1-month free trial

Dealing with outliers: Should cases be removed?

- [Instructor] Sometimes you just have to consider deleting cases that have undue influence on the regression line or that are outliers. But let's talk through the logic of that and use the waste dataset as a demonstration. So I'm gonna go into our resources and launch the waste dataset. Okay, now the waste dataset has severe multicollinearity, so its behavior can be somewhat erratic as a result, but let's take a look. Go ahead and do analyze, regression, linear. I'm gonna purposefully load all five independent variables, even though we know from our familiarity with this dataset that that's probably gonna cause problems, and under save, I'm gonna request Cook's, studentized deleted residuals, DfBetas, and DfFit. Let's go ahead and click on continue and okay. Now, remember that when we request outlier diagnostics, they're gonna show up in the data window, not in the output window. So let's right-click on Cook's, sort descending, and what we see is that there's two cases in particular…

Contents