Associate Professor Dr. Max Buot will present a clever way to use residual plots:
When used as a diagnostic tool in regression analysis, a residual plot (i.e., a scatterplot of predictions versus residuals) should be boring: if points appear below and above y=0 with no discernible trend, then a linear model is an appropriate statistical summary of relationships within the data. In this seminar, we attempt to liven up residual plots and have some fun!
This event will take place on Wednesday, April 24 at 1:30pm in G27 Smith Hall. Light refreshments will be available 15 minutes prior to the start of seminar.