update to schedule and notes.

nickreich · Nov 12, 2018 · 882016b · 882016b
1 parent 5ed563d
commit 882016b
Show file tree

Hide file tree

Showing 4 changed files with 81 additions and 25 deletions.
diff --git a/assets/lectures/lecture8-confidence/lecture8-inference.Rnw b/assets/lectures/lecture8-confidence/lecture8-inference.Rnw
@@ -86,15 +86,9 @@ Relationships are characterized by parameters in our models.
 
 \ 
 
-These are some tools we use to characterize our uncertainty about these parameters:
-\bi
-  \myitem Likelihood
-  \myitem Posterior distributions
-  \myitem Null distributions
-  \myitem Sampling distributions (``classical'' approach)
-\ei
+e.g. in this model, $\beta_1$ characterizes an assumed relationship between FEV and height
 
-There is no one "best" approach. Some will be right for some circumstances, not right for others. Each make different assumptions.
+$$ \widehat{FEV} = \beta_0 +\beta_1\cdot height $$
 
 \end{frame}
 
@@ -125,7 +119,24 @@ There is no one "best" approach. Some will be right for some circumstances, not
 \end{frame}
 
 
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+\begin{frame}[fragile]{Likelihood for coin-flipping}
 
+What is the relative likelihood of each $p$ given 10 coin flips and 5 heads? 
+
+$$L(p | X=5, n=10) = {10 \choose 5}\cdot p^5 \cdot (1-p^{10-5})$$
+
+<<echo=FALSE, fig.height=3>>=
+binom_lik_scaled <- function(p, x, n) {
+  dbinom(x=x, size = n, prob = p)/dbinom(x=x, size=n, p=x/n)
+}
+par(mar=c(5, 5, 1, 1))
+curve(binom_lik_scaled(x, 5, 10), xlab="p", ylab="scaled likelihood", col="blue") 
+@
+
+
+\end{frame}
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
 \begin{frame}[fragile]{Likelihood for coin-flipping}
@@ -172,23 +183,33 @@ Most (not all) statisticians use likelihood to translate data into knowledge.
 A Bayesian incorporates prior belief into the likelihood. What is your prior belief about what this parameter is? 
 
 \bi
-  \myitem Based on prior scientific studies or observations. e.g. "The laws of physics dictate that this coin is more or less fair, so I think the probability of getting a head should be about 0.5." 
-  \myitem Little or no knowledge can be described as having a {\color{orange}uniform prior}. In this case, Bayesian inference is equivalent to just using the likelihood. 
+  \myitem Based on prior scientific studies or observations. 
+
+  e.g.``The laws of physics dictate that this coin is more or less fair, so I think the probability of getting a head should be about 0.5." or ``Based on prior studies, there should be a positive association between height and FEV.''
+
+  \myitem Little knowledge can be described as having a {\color{orange}weakly informative prior}. 
+
+  e.g. ``Based on common sense, we know that height likely has a moderate effect of FEV, but it could be positive or negative, and we don't really want to make an assumption one way or another.''
+
+  \myitem No knowledge can be described as having a {\color{orange}uniform prior}. 
+
+  e.g. ``I have no idea what the relationship between FEV and height should be. it could be anywhere between $-\infty$ and $\infty$.''
+
 \ei
 
 \end{frame}
 
 
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
 
-\begin{frame}[fragile]{Let's establish the classes prior beliefs}
+\begin{frame}[fragile]{Let's establish the classes prior beliefs about this coin}
 
-Go here to submit your guess: https://goo.gl/RKW8dJ
+Go here to submit your guess: https://goo.gl/forms/bp6PE4ZHXedoJhj92
 
 We are going to use these guesses to create a prior distribution for the collective belief in the class about this coin.
 
 <<eval=FALSE>>=
-probs <- read.csv(file="https://goo.gl/jNDbrv")
+probs <- read.csv(file="https://docs.google.com/spreadsheets/d/e/2PACX-1vTd9GsPDSeBtk5Qfx-2xJ-9vSlIaOF7rNYIbVcgWVnx5UC_yoOnQRoQW6kQTMP5kMlPpNaCqrC8m1Ec/pub?gid=1602296264&single=true&output=csv")
 (probs$prob)
 
 library(MASS)

diff --git a/assets/lectures/lecture8-confidence/lecture8-inference.pdf b/assets/lectures/lecture8-confidence/lecture8-inference.pdf
diff --git a/pages/instructor-schedule.md b/pages/instructor-schedule.md
@@ -157,25 +157,60 @@ _Thurs, week 6_
 
 _Tues, week 7_
 
- - Bayesian coin-flipping activity
- - Lecture 8: Understanding uncertainty: coin-flipping, election simulations, etc...
+Project
 
 _Thurs, week 7_
-
- - Lecture 9: understanding uncertainty - sampling distributions, etc...
- - with activities
+
+Project
 
 _Tues, week 8_
 
- - lab 7 (ipod shuffle) in class
+Project
 
 _Thurs, week 8_
 
+Project
+
+_Tues, week 9_
+
+Project
+
+_Thurs, week 9_
+
+Project
+
+_Tues, week 10_
+
+Project
+
+_Thurs, week 10_
+
+Project
+
+
+_Tues, week 11_
+
+ - Biostats pitch
+ - ReichLab research
+ - Lecture 8: lecture8-inference
+
+_Thurs, week 11_
+
+ - Activity from lecture 8
+
+_Tues, week 12_
+
+ - Lecture 9: understanding uncertainty - sampling distributions, etc...
+ - Lecture 9 activity
+
+_Thurs, week 12_
+
  - lab 7 (con't)
 
+<!-- unedited below here -->
+
  _Tues, week 9_
 
- - announcements: career info, survey methods info session
  - hand out old version of exam and discuss differences 
     - logistic regression is fair game this year
     - we have emphasized less the model syntax

diff --git a/pages/schedule.md b/pages/schedule.md
@@ -43,14 +43,14 @@ Month &nbsp; | Day   |     | Topic |  Notes &nbsp;   | HW Due &nbsp;  | Reading
      | Th,  8|     |    | | | 
      | Fri, 9|     |    | | [CC4](../assets/challenges/coding-challenge-4.pdf) |
      | Mo, 12|     | Projects due   | | | 
-     | Tu, 13|     | Understanding Uncertainty | <!--[Lec 8](../assets/lectures/lecture8-confidence/lecture8-inference.pdf), [Suppl](../assets/lectures/lecture8-confidence/lecture8-activity.html) -->| |  Kaplan 11-13
-     | Th, 15|     | Understanding Uncertainty |  <!--[Lec 9](../assets/lectures/lecture9-intervals/lecture9-intervals.pdf), [Suppl](../assets/lectures/lecture9-intervals/lecture9-activity.Rmd)-->  | ||
-     | Fri, 16|    |          | | |
+     | Tu, 13|     | Understanding Uncertainty | [Lec 8](../assets/lectures/lecture8-confidence/lecture8-inference.pdf) | |  Kaplan 11-13
+     | Th, 15|     | Understanding Uncertainty | [Suppl](../assets/lectures/lecture8-confidence/lecture8-activity.html) | ||
+     | Fri, 16|    |          | |  |
      | Tu, 20|     | _Thanksgiving_   | | |
      | Th, 22|     | _Thanksgiving_    | | | 
-     | Tu, 27|     | Understanding Uncertainty | || Kaplan 14-15 
+     | Tu, 27|     | Understanding Uncertainty |   [Lec 9](../assets/lectures/lecture9-intervals/lecture9-intervals.pdf), [Suppl](../assets/lectures/lecture9-intervals/lecture9-activity.Rmd)  |  | Kaplan 14-15 
      | Th, 29|     | Understanding Uncertainty | || Kaplan 14-15
-     | Fri, 30|    |          | | Lab 5 |
+     | Fri, 30|    |          | |  Lab 5<!-- [Lab 5](../assets/labs/lab5-logit/lab5-logit.pdf)-->  |
  Dec | Tu,  4|     | Professional Ethics | | | [MDSR 6](http://mdsr-book.github.io/excerpts/mdsr-ethics.pdf)
      | Th,  6|     | Final exam review | | |
      | Fri, 7|     |     | | Lab 6<!--(../assets/labs/lab6-infer/lab6-infer.pdf)-->|