Statistics 2450, Winter 2020

Outline
Evaluation
Learning center hours
Calendar
Lectures
Assignments
Midterm Exam
Final Exam
Recommanded books
Supplementary Material
Waiting list for registration

Current announcements

Welcome in week #3
This week, we will use two methods (analytical and bootstrap) to see 'if the rat is in the box'.
Assignment #2 is available online. Knit the pdf, print it and give it to me next Monday (Jan. 27 2020)
A help file for assignment #2 is available at
An activity on bootstrapping applied to genetics has been posted.
Welcome in week #4
Theme of this week: 'Sorry, we regress!'
An activity on regression applied to physics has been posted.
Next assignment due Monday Feb 3rd 2020
Typos have been corrected in assignment #3. Please download the last version (29 jan. 2020)
(29/01/20)Today we continue to regress, but we use regression tree instead of polynomial regression
I plan to place a list of exercises online to help you prepare for the midterm
(29/01/20) Added a help file detailing the R code used in today's lecture ( R-mechanics: building regression trees)
Welcome in week #5
On Monday: classification tree, on Wednesday: (changed to:) Practice session
Prepare the midterm: Exercises (MUST DO!)
(Week #5 Feb 3): Assignment #4 is online (Rmd)
Wednesday: we will practice R. This will help you for your assignment 4.
(Week #6 Feb 10):
Monday: Prepare midterm - R programming Lecture (Rmd)
Wednesday: Lecture (Rmd). This will help you for your assignment 5.
Next assignment due Monday Feb 24th 2020: Assignment #5 (Rmd)NOW ONLINE (19/02)
(Week #7 Feb 24):
Monday: Lecture 13 (classification trees) (Rmd) + Prep. Assignment #6
Wednesday: Lecture 14: Combinatorics of the bootstrap (Rmd)
(Week #8 Mar 2):
Next assignment due Monday Mar 9th 2020: Assignment #6 (Rmd)NOW ONLINE (01/03)
Monday: Lecture 15 (Cross-Validation) (Rmd)
Wednesday: Lecture 16: Polynomial regression and bootstrapping
(Week #9 Mar 9):
Next assignment due Monday Mar 23rd 2020: Assignment #7 (Rmd)
Monday: Lecture 17 (Bagging, Random Forest, Boosting) (Rmd)
Wednesday: Lecture 18 (Random Forest, OOB error) (Rmd)
(Week #10 Mar 16):
Next assignment due Monday Mar 23rd 2020: Assignment #7 (Rmd, UPDATED)
Monday: Lecture 19 (Image classification in R) (Rmd)
Added an introductory lecture on Boosting (not required, supplementary material: (pdf))
Wednesday: Lecture 20 (TBA) (Rmd)
(Week #11 Mar 23):
Monday: Revised Syllabus sent to the class
The Final Exam is cancelled.
Monday: Lecture 21 (Image classification in R) (Rmd)
Wednesday: Lecture 22 (html)
Assignment 8 due Monday Apr 6 2020: Assignment #8 (Rmd)
(Week #12 Mar 30):
Monday: Posted on Brightspace: new version of helper file for Assignment 8
Monday: Lecture 23 (html)
Wednesday: Lecture 24 (html)

Outline

Course outline (pdf)

Office Hours: Monday, 1:50-2:15, 4:00-5:00 | Wednesday, 1:50-2:15 | Tuesday, 11:15-12:15

Evaluation

8 assignments + midterm + final. See thecourse outline for details. DUE TO THE COVID-19, the Final Exam is cancelled. The final note will be based on the 8 assignments and the midterm. I have sent a revised syllabus with all relevant details to the class.

The graders for this course are:

Mengyao Wang (email: mn427134@dal.ca)
Arman Kerimbek (email: ar552134@dal.ca)

Please note that Mengyao is also in charge of helping students for STAT240 at the Learning Center (see that section)

Learning centers hours

Math/Stat Learning Centre hours for Stat2450

Location: Chase 119

First day: Tuesday, Jan. 6, 2020

Last day: Friday, April 24, 2020

LC Hours for STAT 2450:

Due to the Covid-19 situation, the information below is not relevant anymore

Mondays Mon. 2-5pm

Tuesdays Tue. 3-4pm ( Mengyao Wang, email: mn427134@dal.ca)

Wednesdays Wed. 2-5pm

Thursdays 3-5pm ( Mengyao Wang, email: mn427134@dal.ca)

Fridays 12-5pm

Closed on the holidays: Feb. 7 (Monroe Day), Feb. 17 (Nova Scotia Heritage Day), Apr.10 (Good Friday)

(Class closed during the study break: Feb 18-Feb 21) Week: Monday-17 off: heritage day

Important dates for 2020

(Dal URL)

Classes Begin January 6, 2020

Munro Day (university closed) February 7, 2020 (OK Friday: not imn 2450)

Nova Scotia Heritage Day (university closed) February 17, 2020 (Monday:first day of study break week) OK

Study Break (no classes – university remains open) February 18 – 21, 2020 OK

Classes End April 3, 2020 (Friday) OK

“In Lieu” Class Day #1 (follows a Friday class schedule) April 6, 2020

Good Friday (university closed) April 10, 2020 OK

Exam Period April 8 – 24, 2020

Spring Convocation (TBC) top

Course Calendar

There will be 24 sessions total (12 bi-weekly lectures).

January: 8 sessions
- Monday, Jan 6, L1
- Wednesday, Jan 8, L2
- Monday, Jan 13,L3
- Wednesday, Jan 15,L4
- Monday, Jan 20, L5, Assignment A1 due
- Wednesday, Jan 22, L6
- Monday, Jan 27,L7, Assignment A2 due
- Wednesday, Jan 29,L8
-
February: 6 sessions
- Monday, Feb 3, L9, Assignment A3 due
- Wednesday, Feb 5,L10
- Monday, Feb 10,L11, Assignment A4 due
- Wednesday, Feb 12,L12
- Monday, Feb 17 BREAK
- Wednesday, Feb 19 BREAK
- Monday, Feb 24, L13,Assignment A5 due
- Wednesday, Feb 26,L14
-
March: 9 sessions
- Monday, Mar 2,L15
- Wednesday, Mar 4,L16
- Monday, Mar 9,L17, Assignment A6 due
- Wednesday, Mar 11,L18
- Monday, Mar 16,L19
- Wednesday, Mar 18,L20
- Monday, Mar 23, L21, Assignment A7 due
- Wednesday, Mar 25,L22,
- Monday, Mar 30,L23
- Wednesday
-
April
- Apr 1, Lecture
- Apr 6 Assignment A8 due

top

Lecture Notes (under construction)

Preliminary lecture (self-learning)
Notes on installing R, Rstudio, RMarkdown, and a quick tutorial intro to R:
(html)
RMarkdown code used to produce this html file:
(Rmd)

Lecture #1: Programming with R - part 1
pdf /Rmd

R as calculator
Data structures
IO
Get help/doc
Workspace
wd

comments
packages
Lecture #2: Programming with R - part 2
pdf /Rmd

Flow controls
Functions
R for statistics

Lecture #3: Programming with R - part 3
pdf /Rmd

User-defined functions
General syntax
call a function within a function
argument
recursion
Lecture #4: Programming with R - part 4
pdf /Rmd

vectorized arithmetics
pdf,auc,percentiles,random sampling
functions of matrices
binom,pois,exp,unif...
d,p,q,r

Lecture #5: Simulation study of the coverage of the t-interval
html Rmd
Details: html Rmd

empirical coverage
simulation
t-interval
plot
Lecture #6: Bootstrap sampling
html /Rmd

bootstrap sample
bootstrap distribution
bootstrap CI
Assess true coverage of CI

Lecture #7: Polynomial regression
html /Rmd

linear regression
adding higher order
bootstrap standard error of slope
Lecture #8: Section 8.1 Fitting a regression tree
html /Rmd

Regression trees
Baseball players salary
cutpoints
bagging
random forests

Lecture #9: Fitting a classification tree
html /Rmd

Classification trees
Predict AHD
Pruning
Assess misclassification rate
Lecture #10: Practice: R programming-1
html/Rmd

matrices
loops
functions
statistical functions
confidence intervals

Lecture #11: Practice: R programming-2
html/Rmd

matrices
loops
functions
Lecture #12: Practice: bootstrapping
html /Rmd

Review of bootstrap
bootstrap sample
bootstrap distribution
bootstrap CI
Assess true coverage of CI

Lecture #13: Classification tree (2)
Rmd

Classification trees
Predict AHD
Pruning
Assess misclassification rate
Lecture #14: Combinatorics of the bootstrap
Rmd

Exchangeable statistics
Counting unique bootstrap patterns
Average fraction missing
Stirling's formula

Lecture #15: Cross-validation
html /Rmd

Classification trees
Cross-validation
Pruning
Assess misclassification rate
Lecture #16: Polynomial regression and bootstrapping
html /Rmd

Bootstrapping with boot
Prepare for assignment 6
Polynomial regression and variants
Review real data examples

Lecture #17:
html /Rmd

Bagging
Random Forest
Regression
MSE
Lecture #18:
html /Rmd

Bagging
Random Forest
Out-of-bag
Titanic

Lecture #19:
html /Rmd

Image classification
Random Forest
MNIST data
Training set
Testing set
Lecture #20: Cancelled
html /Rmd

Chat on Brightspace

Lecture #21:
html /Rmd

Image classification
Alternative models
MNIST data
Training set
Testing set
Lecture #22: Preparation for Assignment 8

Wednesday:
html

Properties of the covariance
Review of validation procedures
Chat on Brightspace
Boosting is (unrequired yet interesting) supplementary material

Lecture #23: Preparation for Assignment 8

Monday:
html

Tests of hypotheses and p-values
Plotting details
CART and random forests
Chat on Brightspace

Lecture #24: Preparation for Assignment 8

Wednesday:
html

Final revisions
Chat on Brightspace

top

Assignments

Assignment # 1, due January 20th Rmd R calculator IO with scan shaping arrays read csv boxplot	Assignment # 2, due January 27th Rmd help /sol read csv edit data binning scatterplot manipulate booleans create dataframe boxplot
Assignment # 3, due February 3rd Rmd For loop If then Modify code Random matrix	Assignment # 4, due February 10th Rmd write functions obeying goals specify arguments specify returned values t-Confidence interval
Assignment # 5, due February 24th Rmd (ONLINE) generate random observations CI for mean bootstrap sampling bootstrap medians recursive function	Assignment # 6, due March 9th Rmd bootstrap and linear regression bootstrap of slope estimate histogram
Assignment # 7, due March 23rd< /Rmd fit tree to training data cross-validation pruning testing error	Assignment # 8, due April 6 /Rmd 6 problems (total 50 points)

top

Midterm Exam

Midterm is February 10th, 6:00-8:00 PM

Location: D420 (auditorium 1), McCain Building, University Avenue top

Final Exam

Cancelled due to COVID-19-related measures.

Recommanded book

top

Introduction to Statistical Learning by James Gareth et al.

Supplementary material

top

(Week #2, Jan 14): Help for Assignment #1
(Week #2, Jan 14): Hey, let's talk about distributions!
(Week #2, Jan 16): Hey, let's practice common distributions in R!
(Week #2, Jan 16): Hey, let's practice matrices in R!
(Week #3, Jan 20): The speed of light and its uncertainties
(Week #3, Jan 22): Uncertain phylogenies
(Week #4 Jan 27): Van der Waals'idea was not a regression
(Week #4 Jan 29): R-mechanics: building regression trees (Rmd)
(Week #4 Jan 29): Writing maths in Rmd files (Rmd)
(Week #4 Jan 31): I have revised and enriched the tree.Rmd file
(Week $5 Feb 3): Exercises for the midterm (Rmd)
(Week $10 Mar 13): Introduction to Boosting (pdf)

top

Waiting list for registration

The number of students that can attend the course is limited by room capacity (120).

I try my best to accommodate all students who want to take this class. Please note that I made a first set of changes on Jan. 7 2020 to allow students then in the waiting list..

I have made further changes today (Jan. 14) to allow more students in.
Please note that you have to register yourself, following my allowing you to take the course.

orests: who and why
html -->

Lecture #1: Programming with R - part 1 pdf /Rmd R as calculator Data structures IO Get help/doc Workspace wd comments packages	Lecture #2: Programming with R - part 2 pdf /Rmd Flow controls Functions R for statistics
Lecture #3: Programming with R - part 3 pdf /Rmd User-defined functions General syntax call a function within a function argument recursion	Lecture #4: Programming with R - part 4 pdf /Rmd vectorized arithmetics pdf,auc,percentiles,random sampling functions of matrices binom,pois,exp,unif... d,p,q,r
Lecture #5: Simulation study of the coverage of the t-interval html Rmd Details: html Rmd empirical coverage simulation t-interval plot	Lecture #6: Bootstrap sampling html /Rmd bootstrap sample bootstrap distribution bootstrap CI Assess true coverage of CI
Lecture #7: Polynomial regression html /Rmd linear regression adding higher order bootstrap standard error of slope	Lecture #8: Section 8.1 Fitting a regression tree html /Rmd Regression trees Baseball players salary cutpoints bagging random forests
Lecture #9: Fitting a classification tree html /Rmd Classification trees Predict AHD Pruning Assess misclassification rate	Lecture #10: Practice: R programming-1 html/Rmd matrices loops functions statistical functions confidence intervals
Lecture #11: Practice: R programming-2 html/Rmd matrices loops functions	Lecture #12: Practice: bootstrapping html /Rmd Review of bootstrap bootstrap sample bootstrap distribution bootstrap CI Assess true coverage of CI
Lecture #13: Classification tree (2) Rmd Classification trees Predict AHD Pruning Assess misclassification rate	Lecture #14: Combinatorics of the bootstrap Rmd Exchangeable statistics Counting unique bootstrap patterns Average fraction missing Stirling's formula
Lecture #15: Cross-validation html /Rmd Classification trees Cross-validation Pruning Assess misclassification rate	Lecture #16: Polynomial regression and bootstrapping html /Rmd Bootstrapping with boot Prepare for assignment 6 Polynomial regression and variants Review real data examples
Lecture #17: html /Rmd Bagging Random Forest Regression MSE	Lecture #18: html /Rmd Bagging Random Forest Out-of-bag Titanic
Lecture #19: html /Rmd Image classification Random Forest MNIST data Training set Testing set	Lecture #20: Cancelled html /Rmd Chat on Brightspace
Lecture #21: html /Rmd Image classification Alternative models MNIST data Training set Testing set	Lecture #22: Preparation for Assignment 8 Wednesday: html Properties of the covariance Review of validation procedures Chat on Brightspace Boosting is (unrequired yet interesting) supplementary material
Lecture #23: Preparation for Assignment 8 Monday: html Tests of hypotheses and p-values Plotting details CART and random forests Chat on Brightspace	Lecture #24: Preparation for Assignment 8 Wednesday: html Final revisions Chat on Brightspace