This week, we continue our study of parametric probability models by talking about the bivariate normal distribution, and about the stock market.
Topics for Monday:
- last week’s homework (see kelly.R).
- regression to the mean
Topics for Wednesday:
- the bivariate normal distribution
- the bivariate normal model for stocks and bonds.
- (if time) Monte Carlo simulation of stock/bond portfolios
Reading
Please read the rest of Chapter 15 on the bivariate normal distribution. You can safely skip the section entitled “Functions of random variables (advanced topic).” You are not responsible for this section in this course.
I have also posted a short (2.5 page) set of summary notes on bootstrap resampling, the small lecture topic on Wednesday.
For Monday of next week, please read through the middle of page 305 of Chapter 16. You can safely stop once you reach the section entitled “Advanced topic: estimating the parameters of the logistic regression model.” You are not responsible for anything beyond this section.
Optionally, you might be interested in professional poker player Phil Laak’s take on the Kelly criterion.
Software
For Monday, please download the R script bivariate_normal.R from the R Scripts tab above. For the bivariate normal script to work, you will need to first install the mvtnorm library in R.
Finally, if you want to get ahead for next week, you can look to the R walkthrough on logistic regression. I will ask you to complete this by the time you come to class next Wednesday (last class day of the semester).
Exercises
This week marks the last set of exercises for the semester. Exercises 9 are on simulating portfolios and logistic regression models. They are due by 5 PM on the last day of the semester: Friday, May 4, 2017. Please submit them as a PDF – not a Word file – to statdropbox@gmail.com (same e-mail as for projects).