When to use regression on SAT Math in Desmos
- Topic
- Regression
- By
- SolvePlot
Use regression when a question gives data or constraints for a model with unknown parameters. Enter the data, use a tilde (~) to fit the stated model, and evaluate the quantity requested. A fit is only useful if the model and its assumptions match the question.
An equals sign defines an equation. A tilde asks Desmos to fit free parameters to data. For a quadratic model, three distinct input values can determine its three coefficients when the data are consistent.
Fitting three points does not establish that an unknown real-world relationship is quadratic. In the example below, the question explicitly tells us that the function is quadratic.
How do you fit a quadratic from a table?
Keep the input and output values in matching order. You can use a table, or enter two lists called and The quadratic model uses a, and c as the parameters to fit.
After the fit, define a function using those parameters and evaluate it. This lets Desmos do the arithmetic while keeping the connection to the original model visible.
Original example
Work through the question
A quadratic function has values and What is
AnswerWhat to enter in Desmos
Use one expression row for each entry. Underscores create subscripts; the tilde ~ fits parameters.
x_1=[0,1,2]Enter the three inputs in the order given. The underscore makes the subscript.
y_1=[3,6,11]Enter the corresponding outputs in the same order.
y_1~ax_1^2+bx_1+cUse a tilde, not an equals sign. Desmos fits a = 1, b = 2, and c = 3.
f(x)=ax^2+bx+cDefine the fitted quadratic as a function.
f(4)Evaluate the requested input. The result is 27.
Read the result
Read 27 beside f(4). The fitted curve passes through all three data points. The three fitted coefficients describe this specified quadratic; do not round them before evaluating the function.
The fitted function is It returns and at inputs and At it gives
The algebraic solution
- Write Since
- The other values give and Subtract twice the first equation from the second to get so and
- Then
When is each method useful?
Use Desmos when
Regression avoids solving for the coefficients by hand when a model is given and enough independent data determine it. It is especially helpful with less convenient values.
Use algebra when
Algebra may be preferable when the question asks for an exact symbolic relationship, a proof, or why a parameter is not determined. A model with too many free parameters can fit the givens without fixing the requested answer.
Common mistakes to avoid
- Start with a blank graph so a, b, and c are free parameters, rather than values left over from another problem.
- Use ~ for a fit. Using = may graph a relationship or create sliders instead.
- Keep x_1 paired with y_1; reversing just one list changes the data.
- Use the model named by the problem. A good-looking fit is not evidence that you chose the right model.
- A fit to approximate data is approximate. Do not treat a rounded regression coefficient as an exact proof.
Try one yourself
A quadratic has and Find
See the answer and steps
The fitted model is so The three given outputs agree with this model.
Desmos entries:
x_1=[0,1,2]
y_1=[2,5,10]
y_1~ax_1^2+bx_1+c
g(x)=ax^2+bx+c
g(3)Calculator references
The examples above are original. These official resources explain the calculator features and test-day rules.