Exercises (afternoon - 24 April)

broom and modelr

Exercises:

• In the final R chunk in day04_t01_manymodels.html the tidy function is used to extract the coefficients. Set quick = FALSE and conf.int = TRUE and include the 95% confidence intervals for (Intercept) and year. Make sure that the output reads easily.
• 1, 2 and 3 in Section 25.2.5 of R for Data Science

Cross-validation and Bootstrap

Look at the R lab Sections 5.3 on Cross-validation and Bootstrap of Introduction to Statistical Learning with applications in R (ISLR) by James, Witten, Hastie and Tibshirani. Freely available from statlearning.com.

Exercises:

• 8 and 9 (pp. 200 - 201) from Section 5.4 in ISLR
• The exercise below.

A simple model

The ToothGrowth data set (see ?ToothGrowth for explanation) contains observations on 60 guinea pigs treated with either vitamins and or orange juice.

• Fit a linear model to the data. Let dose be numeric.
• Investigate if there is an interaction between dose and supp.

Is the interaction really significant?

• Compute the bootstrap confidence intervals. Both by case and residual bootstrap re-sampling.
• Does your conclusion change?
• Which one is preferable in for this type of data?