Notes: (Topics 1.13 – 1.
14) Modeling Functions and Applications Solutions
The idea of modeling (also called regression) is a major component of the AP Precalculus curriculum and will be
revisited throughout the course. These topics will utilize the graphing calculators heavily and often include context.
In Unit 1, we will work with regression models for polynomial functions (linear, quadratic, cubic, and quartic) and for
rational functions.
Building Regression Models on the Graphing Calculator
Building a regression model on the graphing requires two steps:
1. Entering the data to be modeled
2. Selecting the regression model
Step 1: Press the “stat” button on the TI – 84 and select “1: Edit…” from the menu.
Enter the data into the lists with L1 = 𝑥 and L2 = 𝑦
Step 2: In the “stat” menu, arrow to the right to the “CALC” menu.
Select the desired regression model.
Note: There are 2 linear regression options: 4: LinReg (𝐚𝒙 + 𝐛) and 8: LinReg (𝐚 + 𝐛𝒙).
These two options are essentially equivalent – consider option 8 (AP Stats)
Use 𝐋𝟏 for Xlist: and 𝐋𝟐 for Ylist:
(𝐋𝟏 and 𝐋𝟐 are in blue above the “1” and “2” buttons: Press “2nd” and then “1” or “2”)
Tip: For Store RegEQ:, enter 𝐘𝟏. (Press “alpha” and then “trace” to select 𝐘𝟏)
Notes: Function Models and Applications Topics 1.13 – 1.14 Created by Bryan Passwater
𝑡
4 5 6 8 12
(age in weeks)
𝑊(𝑡)
4.2 4.4 4.8 5.1 5.7
(weight in kg)
Example 1: The age (in weeks) and weight (in kilograms) of 5 randomly selected babies from a particular pediatrician’s
office are listed in the table above.
A linear regression 𝑦 = 𝑎 + 𝑏𝑥 can be used to model these data, where 𝑦 is the predicted weight of a baby (in kg) that is
𝑥 weeks old.
a) Write the equation of the linear model for these data.
Enter data in L1 and L2 Select 8: LinReg (a + b𝑥) Enter L1, L2 and Y! Get regression equation
𝐋𝐢𝐧𝐞𝐚𝐫 𝐑𝐞𝐠𝐫𝐞𝐬𝐬𝐢𝐨𝐧 𝐌𝐨𝐝𝐞𝐥: 𝑦 = 3.545 + 0.185𝑡
b) Using the linear model from part a), what is the predicted weight (in kilograms) of a baby that is 10 weeks old?
Tip: You can quickly find values of
functions that are saved in Y! from the
home screen on the calculator.
c) The weight of a sixth baby is 5.3 kg. Using the model from part a), what is the age (in weeks) of this baby?
5.3 − 3.545
3.545 + 0.185𝑡 = 5.3 ⇒ 𝑡 = = 9.4864 … ≈ 9.5 weeks
0.185
Notes: Function Models and Applications Topics 1.13 – 1.14 Created by Bryan Passwater
Residuals
When we use a model to predict values, we expect our model to produce values reasonably close to the actual values, but
our models are not expected to result in exact values generally. The difference between an actual value and the value
predicted by a model is called a residual.
𝐑𝐞𝐬𝐢𝐝𝐮𝐚𝐥𝐬
Residual = Actual Value − Predicted Value
Residual = 𝑦 − 𝑦\
𝑡
4 5 6 8 12
(age in weeks)
𝑊
4.2 4.4 4.8 5.1 5.7
(weight in kg)
Example 2: Using the model found in Example 1, what is the residual of the baby that is 5 weeks old? Interpret the
meaning of this value in the context of this problem.
Actual Value: 4.4
Predicted Value: 𝑦\ = 3.545 + 0.185(5) = 4.47
Residual: 4.4 − 4.47 = −0.07
Interpretation: The predicted value is 0.07 weeks larger than the actual value 4.4 weeks.
Example 3: Using the model found in Example 1, a baby that was 7.5 weeks old had a residual of -0.7 kg. What was
the actual weight (in kg) of the 7.5 week old baby?
Residual = 𝑊 − [3.545 + 0.185(7.5)] = −0.7 ⇒ actual weight 𝑊 = 4.9325 − 0.7 = 4.2325 kg
Example 4: Sketch the residual plot for the five data points in the table.
t W 𝑦\ resid
0.1 4 4.2 4.285 -0.085
5 4.4 4.47 -0.07
6 4.8 4.655 0.145
8 5.1 5.025 0.075
–0.1 12 5.7 5.765 -0.065
4 5 6 12
Notes: Function Models and Applications Topics 1.13 – 1.14 Created by Bryan Passwater
Residual Plots on the TI-84
Important: To create a residual plot, you must have already entered the data and calculated the regression model.
Step 1: You will need to press the “2nd” and then the “𝒚 =” buttons to open the “statplot” menu.
Turn the plot “On” and
Choose “1: Plot1…Off”
select the first plot type.
Step 2: We need “Xlist: L1” and “Ylist: RESID”, so we need to change “Ylist:” from “L2” to “RESID”.
Arrow down to “Ylist:” and
The “Plot1” options should
press “2nd” and “stat”.
look like this:
Select “7: RESID”
Step 3: Graph the residual plot: Press “Zoom” and then “9: ZoomStat”.
“Zoom” is a great to help create Select “9: ZoomStat” when
Our residual plot:
an appropriate viewing window. working with data.
Notes: Function Models and Applications Topics 1.13 – 1.14 Created by Bryan Passwater
𝑥 0 0.4 0.9 1.2 1.7 2.2 2.9 3.4
𝑦 5 10.6 15.4 17.1 18.0 16.5 10.2 2.8
Example 5: The table above provides data for 8 ordered pairs (𝑥, 𝑦).
a) Which function type best models the data in the table: linear, quadratic, cubic, or rational? Explain your answer using
characteristics from the data in the table. The data has characteristics of a quadratic model. The y–values increase then
decrease and there is an approximate symmetry about 𝑥 = 1.5. See the scatterplot below
b) Write the equation of the regression model for the data in the table. 𝑦 = −4.8831 … 𝑥 " + 15.95804 … 𝑥 + 4.9956 …
x 0.3 1.6 3.2 4.7 7.1 8.8 10.3
y 0.9 3.3 4.9 5.7 6.5 7.0 7.1
Example 6: The table above provides data for 7 ordered pairs ( x, y ) .
a) (i) Using a logarithmic regression, construct a natural logarithmic function model f for y as a function of x .
Write an expression for f ( x ) in the form f ( x ) = a + b ln x , where a and b are constants.
2.8636 … + 1.8184 ln 𝑥
(ii) The residual plot is provided below. What are the coordinates of the labeled point R on the residual plot?
(iii) Does the residual for the point R indicate that the model f underestimates or overestimates the value of y at
this x-value ? Use the residual value to explain your reasoning. 𝑅 = (1.6, −0.4183 … ) R is negative which
means the Residual = 𝑦 − 𝑦\ < 0 ⇒ 𝑦\ > 𝑦, so 𝑓(1.6) overestimates the data value 3.3 by 0.4183...
Notes: Function Models and Applications Topics 1.13 – 1.14 Created by Bryan Passwater
