Studio 6
----------------------------
Topics: Continuous data, Beta priors
R: simulation
---------------------------
Before class:
1. Post studio6.zip containing: studio6.r, studio6-sol.r, studio6-prep.r, studio6-loaddata.r, studio6dataframe.csv, tradesdata0.csv
2. Students should download this zip file and unzip it in their 18.05 R directory.
---------------------------
For class:
1. Use studio6-slides.pdf
2. Follow class script below.
---------------------------
After class:
1. Post studio6-slides-all.pdf
---------------------------
Class script
Slide 1.
Everyone should have downloaded and unzipped studio6.zip
Slide 2. Describe NASDAQ data and project
-- 2 minutes --Just to give a sense of how the data is packaged
Slides 3-6 Exploration
-- to save time for the project this needs to be quick.
-- Main point: you always start with exploration.
-- Hypothesis of exponential distribution for wait times is reasonable
from the exploration and for theoretical reasons.
Slide 7: BOARD QUESTION
This will take some time. The question does not give a lot of details, so
many students will need a lot of prompting.
-- Problem 1: The class will need reminders on the pdf of an
exponential distribution and on continuous-continuous updating
Tell students not to try computing the integrals. This is for the numerics.
-- Problem 2: Look at the code outline on the slide 8. Students will
think they need to give more detail than this.
DISCUSSION:
Problem 1: Show likelihood function and update table.
Don't do any computations.
Problem 2: On next slide
Slides 8, 9: Coding outline
Relate to theoretical answer.
Say it, but don't dwell on normalizing pdf's.
Slide 10: R Project
--Want to leave lots of time for this
--Point out studio6.r has some example code.
IN PARTICULAR, it shows how to load and use studio6dataframe.csv
--They will have solution code in studio6-sol.r.
Try to keep them away from this until they have made a real effort
without it.
Slides 11-14: Plots of solutions
--There probably won't be time to show details of studio6-sol.r, so
just remind them of its existence.
--On the MAP vs time plots the orange curve is a least squares fit
of a parabola. We haven't covered this yet, so the details
can be described as a coming attraction.
Slide 15: Price plot
-- This can be SKIPPED.
-- This is off the main topic. It was available in the data.