Regression finds coefficients for a model such as y = b0 + b1x1 + b2x2 that minimize the sum of squared residuals. It quantifies how much the response moves per unit change in each input, gives prediction intervals and tests which terms matter. With designed experiments the inputs are balanced and the estimates are clean; with historical production data, correlated inputs and lurking variables can make the coefficients misleading.
Run a regression of CD on dose and focus for the last 200 lots and see if the dose term still holds.
Also heard as
- regression analysis
- linear regression
- multiple regression
Related terms
Residual
Difference between an observed value and the value a fitted model predicts for it.
R-squared
Share of the variation in a response that a regression model accounts for, on a scale from 0 to 1.
Leverage
Measure of how far a data point's input values sit from the rest, which gives it outsized pull on a fitted model.
ANOVA
Analysis of variance: splitting total variation into parts from each factor and from error, then testing which parts are real.