Choosing among algebraic model families

Choosing among algebraic model families

Choose and critique linear, quadratic, power and rational models using both data and context. This precalculus practice asks what a proposed formula explains, what restrictions it must satisfy and what further evidence would distinguish it from competing formulas.

Questions include unequal input spacing, constant-area constraints, a square-root inverse relation, bounded growth and two models that agree at existing observations. Residual and validation tasks compare numerical fit with the structure of the errors. A final algebraic counterexample shows why finitely many exact observations do not establish one universal model.

Prerequisites include linear and quadratic models, power functions, rational behavior and the meaning of a residual. All datasets are synthetic and supplied in the questions. Single-select and multiselect items progress from direct model selection to designing a useful comparison and evaluating the limits of the evidence.

Check formulas against observations, but also check units, allowed inputs, expected long-run behavior and the reason a relationship should hold. Detailed explanations separate a best choice among named candidates from a proof about every possible function. Use independent validation evidence and inspect error patterns when supplied; a small training error or an attractive curve alone does not guarantee reliable extrapolation. The set supports reasoned model selection without requiring advanced regression theory.