Welcome to STA248H1 - Statistics for Computer Scientists

Course Instructor:

• Wei Q. Deng
• Office hours: Monday 4-6pm, Wednesday 4-6pm (from second week)
• Office hour location: SS623B (TA office in the basement of Sidney Smith - starting July 9th)

Final exam

• Date: August 17th
• Duration: 9am-12pm
• Location: EX200
• If you want information (no concept nor definitions) included on the final exam information sheet, please email me between July 24th (after midterm) and July 28th, and I will try to accommodate.
  
• Review lecture: August 13th 6-9pm, please bring your questions to class and be prepared to work.
• Extra office hours:
  • August 13th, 2-6pm (SS623B);
  • Cancelled: August 16th, 9am-11am and 12pm-2pm (SS623B);
  • TA office hour August 16th 11am-12pm (SS623B)
• The first and last a few pages of the final exam is here

Lecture Slides

• lecture 9 (August 13th) is a review lecture of all topics covered. Please try as many questions as you can and bring any questions to the class.

  • Partial solution + hints can be found here. Note for the one-sided interval the signs and the size of the interval are now updated.

• lecture 8 (August 8th) covers the introduction to linear regression with a little bit of theory in connection to hypothesis testing as well as confidence intervals.

  • Relevant readings from OpenIntro is Chapter 7; alternatively you can look at Chapter 12 of Probability and Statistics for Engineering and the Sciences.

• lecture 7 covers the introduction to linear regression with connection to supervised learning and examples. The lecture will take 2 hours and covered by your TA Tianle.

  • The python notebook files are here and here.
  • Relevant readings from OpenIntro is Chapter 7; alternatively you can look at Chapter 12 of Probability and Statistics for Engineering and the Sciences.

• Today’s lecture will be a review of midterm solutions (Q3-Q5). Solutions will NOT be posted. You can pick up the midterm from your TA Tianle who will be covering for me Monday.

  • Tests written in pencil will not be remarked after the midterm review.
  • If you plan to request a remark, please write a brief justification for why you deserve the marks.
  • Any request for a remark should be send to me via email by August 5th.

• lecture 6 and the rmarkdown file here cover the remaining topics in hypothesis testing such as power and type I and II error, as well as some simulation examples. We also reviewed the first two questions on the midterm.

  • Note the change on slide page 36 (power calculation for t-test will not be on the final exam).
  • Relevant readings from OpenIntro are Chapters 4.3, 5.4, and 5.5; alternatively you can look at Chapter 8 of Probability and Statistics for Engineering and the Sciences.

• lecture 5 contains a quick review of topics learnt so far and partial solution to some of the questions.

  • Relevant readings from OpenIntro are Chapters 4 and 5; alternatively you can look at Chapters 7, 8, 9 of Probability and Statistics for Engineering and the Sciences.
  • Typo on page 38, Step 2: the second equality has the inequality sign reversed, i.e. \(>\) instead of \(<\) because of multiplying each side by \(-1\).

• lecture 4 and rmarkdown document to generate the slides can be found here for you to look at the R/Python code chunks.

  • There was a typo on slide page 32 for the two-sided test p-value, which has been corrected so it is now consistent with slide page 35.

  • There was a typo on slide page 31 for the one-sample proportion test, the proportion in the denominator should be \(p\) instead of \(\hat{p}\), i.e. taking the parameter value specified by the null hypothesis. The version with the \(\hat{p}\) in the denominator is used when we calculate confidence interval without any known parameter value.

  • Relevant readings from OpenIntro are Chapters 4 and 5; alternatively you can look at Chapters 7, 8, 9 of Probability and Statistics for Engineering and the Sciences.

• lecture 3 and rmarkdown document to generate the slides can be found here for you to look at the R/Python code chunks.

  • Relevant readings from OpenIntro are Chapter 4, specifically 4.1-4.3, and 4.5; alternatively you can look at Chapters 7, 8, 9 of Probability and Statistics for Engineering and the Sciences.

• lecture 2 and rmarkdown document to generate the slides can be found here for you to look at the R/Python code chunks.

  • To practice what you learnt, make sure you know each of the concepts introduced as well as being able to solve the examples and exercises on the lecture slides.
  • Relevant readings from OpenIntro are Chapters 2, 3 (review for STA247) and 4.4; alternatively you can look at Chapters 2-4 (review for ST247) and Chapters 5 and 6 of Probability and Statistics for Engineering and the Sciences.
  • You can also refer to your slides or textbook from STA247 if you need a refresher on concepts such as expectation, variance and moment generating functions.

• lecture 1 and rmarkdown document to generate the slides can be found here for you to look at the code chunks slide page 40 has been updated to reflect a more accurate depiction of a left skewed distribution.

  • To practice what you learnt in Lecture 1, you can explore the car dataset (data description here). For each variable, generate what you think is the best graphical and numerical summaries. You should be able to defend your choices using the material from the lecture slides.
  • Relevant reading material from the OpenIntro Chapter 1; alternatively you can look at Chapter 1 of Probability and Statistics for Engineering and the Sciences.

R and Python programming resources

• R

  • R cookbook
  • R intro from Cran
  • Quick R

• Python

  • statistical analyses
  • pydataframe
  • NumPy and Scipy

DataCamp

You can take a look at DataCamp. The invitation link has been sent out, if you have not received, please let me know.

For students using R, you should complete one of the intro stats modules and one data visualization modules:

• Intro to Statistics with R: Analysis of Variance (ANOVA)
• Intro to Statistics with R: Student’s T-test
• Data Visualization with ggplot2 (Part 1)
• Data Visualization with ggplot2 (Part 2)

For students using Python, you should complete one of the intro stats modules and one data visualization modules:

• Statistical Thinking in Python (Part 2)
• Introduction to Data Visualization with Python

I have added additional modules that you might find helpful to get you started, but you will not be marked based on those.

The deadline for the completion of two modules is set for August 5th.

You can still complete the modules after August 6th - for each additional day after the due date, a 10% penalty will be applied.

Midterm date

• Date: July 23rd
• Duration: 18:00-21:00
• Location: EX300 and EX310
• Due to the high demand for it, here is the first page and last a few pages of the midterm (a.k.a ``formula’’ pages).
• For those of you who missed the midterm due to legitimate reason, please check the faculty’s policies regarding a missed test/exam here and here. Once your absence has been justified, your final exam will count towards 90% of your final grade.
• Midterm marks are now up on portal
• The class average is 46 and class median is 50.5. Here is a parametric bootstrap inspired distribution of a hypothetical class marks assuming the marks are normally distributed and the parameters were estimated from the actual marks.

Class schedule

• Monday 18:00-21:00 in SS 2118 (no lecture on August 6, Civic holiday)
• Wednesday 18:00-21:00 in SS 2118.