Regression

Regression fits a model to data by minimizing an error measure, often the sum of squared residuals. It is useful for noisy engineering measurements.

Governing FormulaSSE = sum of (y_i - y_hat_i) squared

10 Questions

1.What is the primary purpose of regression?

2.Which practice improves reliability when applying regression?

3.What is a common numerical concern in regression?

4.Why are tolerances used in regression algorithms?

5.What should be compared when validating a regression result?

6.What can happen when algorithm settings are poorly selected in regression?

7.Which engineering habit is important when documenting regression?

8.What does convergence usually indicate in regression?

9.Why should physical engineering judgment still be applied to regression results?

10.Which factor can most directly affect the computational cost of regression?