Which term is associated with the residual defined as Actual minus Predicted?

Which term is associated with the residual defined as Actual minus Predicted?

• A. Residual Plots
• B. Cook's Distance
• C. Features
• D. Heteroskedasticity

Answer: (A)

Residuals are the differences between what the model predicted and what actually happened, calculated as Actual minus Predicted. To understand and diagnose these differences, we use residual plots. A residual plot shows the residuals on the vertical axis against fitted values or an input feature on the horizontal axis. If the residuals are scattered randomly around zero with no clear pattern, the model fit is reasonable and the assumptions (like linearity and constant variance) are likely met. If you see patterns—such as curves, funnels, or increasing or decreasing spread—that signals the model may be misspecified, there might be a nonlinear relationship, or the variance of residuals changes with the level of the prediction (heteroskedasticity). Residual plots are the go-to tool for visualizing and diagnosing these issues. The other options refer to separate ideas: Cook's Distance measures how influential a data point is on the regression fit; Features are the input variables; Heteroskedasticity describes non-constant residual variance and is something you look for in residuals, often via the plot, but it’s not the term for the residuals themselves.

Residuals are the differences between what the model predicted and what actually happened, calculated as Actual minus Predicted. To understand and diagnose these differences, we use residual plots. A residual plot shows the residuals on the vertical axis against fitted values or an input feature on the horizontal axis. If the residuals are scattered randomly around zero with no clear pattern, the model fit is reasonable and the assumptions (like linearity and constant variance) are likely met. If you see patterns—such as curves, funnels, or increasing or decreasing spread—that signals the model may be misspecified, there might be a nonlinear relationship, or the variance of residuals changes with the level of the prediction (heteroskedasticity). Residual plots are the go-to tool for visualizing and diagnosing these issues. The other options refer to separate ideas: Cook's Distance measures how influential a data point is on the regression fit; Features are the input variables; Heteroskedasticity describes non-constant residual variance and is something you look for in residuals, often via the plot, but it’s not the term for the residuals themselves.