On Monday in class, we will practice fitting nonlinear curves. To prepare, make sure that you have already watched the videos, completed the software walkthroughts, and read the material on nonlinear equations from last week’s post.

Our in-class case study is on finding an optimal price for milk, given historical sales and price data. This is one of your homework problems this week, and it will finish out Chapter 2 of the course packet.

In the past, some students have found it useful to complete the (optional) mini walkthrough on how to optimize a function in R. This isn’t necessary to do before coming to class.

On Wednesday, we will learn to quantify forecasting error using prediction intervals. A great example of prediction interval is the kind of growth chart that pediatricians use to show the normal range of variability in height and weight for growing babies, as a function of age. Charts for male babies here, charts for female babies here. (Really, CDC? Pink and blue? So gender normative!) You can also find a cool interactive version of the chart here.

We will also learn about the decomposition of variance. This material comes from chapter 3 of the course packet.

Videos

Please watch the following videos to prepare for class next week (week 4):

Software practice

Complete the following software walkthroughs. For Wednesday of this week (week 3) in class:

For Monday of next week (week 4):

  • reaction time in video games: modeling numerical outcomes with more than one categorical predictor; dummy variables and interaction terms; analysis of variance.

Readings

From the course packet, please read Chapter 3 by Wednesday of this week. For Monday of next week (week 4), please read Chapter 4 through page 94 (stopping at the section on “Numerical and grouping variables together.”)

Exercises

The exercises this week are about linear and nonlinear regression. They are due on Monday, February 5 at the beginning of class.