Course outline (2013)

From Statistics for Engineering
Jump to navigation Jump to search
Class date(s): 08 January 2013
Nuvola mimetypes pdf.png (PDF) Course outline
Nuvola mimetypes pdf.png (PDF) Projector overheads

<rst> <rst-options: 'toc' = False/> <rst-options: 'reset-figures' = False/> Logistics

=

    • Instructor**

Kevin Dunn, `kevin.dunn@mcmaster.ca <mailto:kevin.dunn@mcmaster.ca?Subject=4C3%20course>`_ (BSB, room B105)

    • Teaching assistants**

Maryam Emami, Room JHE-361, ext 23263, `emamis2@mcmaster.ca <mailto:emamis2@mcmaster.ca?Subject=4C3%20course&cc=kevin.dunn@mcmaster.ca>`_

Shailesh Patel, Room JHE-369, ext 24031, `patelsr@mcmaster.ca <mailto:patelsr@mcmaster.ca?Subject=4C3%20course&cc=kevin.dunn@mcmaster.ca>`_

    • Class time and location**

T13, room 127. Tuesday, Wednesday and Friday, 12:30 to 13:20. First class: Tuesday, 08 January 2013.

    • Disclaimer**
	This outline **may be modified**, as circumstances change.

About the course

=====
    • Official description**

Linear regression analysis in matrix form, non-linear regression, multi-response estimation, design of experiments including factorial and optimal designs. Special emphasis on methods appropriate to engineering problems.

    • What you must be able to demonstrate by the end of the course**

* Understand that all data has variability: we want to separate that variability into information (knowledge) and error (unknown structure, noise, randomness). * Interpret confidence intervals and univariate data statistics (mean, median, histograms, significant differences). * Understand and use process monitoring charts. * Least-squares models: how to fit and especially how to interpret them, understand the confidence limits and model limitations. * Be able to design your own experimental program and then also interpret experimental data. * Understand the principles of latent variable methods for engineering data.

  • Note:* there are no tutorials scheduled for 4C3/6C3. There is a however a significant chunk of time you will have to invest outside of classes to learn new computer software. Web-based tutorials will be provided for you to complete this self-directed learning in your own time.
    • Prerequisites**

A basic course in statistics that covers probability, means, variances, confidence intervals and linear regression. However, all these topics are covered again in this course, focusing on their practical application to engineering problems.

    • Course materials**

The course website will be permanently available at `http://learnche.mcmaster.ca/4C3 <http://learnche.mcmaster.ca/4C3>`_. You may use this as a resource even after you graduate.

Course materials, assignments and solutions will be available from the website. Course announcements will only be posted to the main page of the website - students are expected to check the website at least 3 times per week.

.. raw:: latex

\newpage

    • Required textbook**

There is no official course textbook. We will be using the instructor's own material from his book, `Process Improvement using Data <http://learnche.mcmaster.ca/pid>`_. The book was written specifically for this course, and will be available as a PDF from the course website. It is your responsibility to print out these notes and bring them to class.

If you prefer not to print it yourself, the *Titles Bookstore* will have a limited number of printed copies, available for only the cost of printing the PDF.

    • Recommended readings**

If you would like to buy one book to supplement the course material, I highly recommend the first one, Box Hunter and Hunter, for its practical engineering perspectives on data analysis.

#. G.E.P. Box, J.S. Hunter, and W.G. Hunter, *Statistics for Experimenters - Design, Innovation and Discovery*, 2nd edition, Wiley. ISBN: 978-0471718130. #. D.C. Montgomery and G.C. Runger, *Applied Statistics and Probability for Engineers*.

Other reference texts are listed on the course website and are generally available in Thode Library.

    • Course outline (differs somewhat from the official description)**

The course is divided into 6 main sections, taught over 12 weeks.

#. *Visualizing data*: creating high-density, efficient graphics that highlight the data honestly. #. *Univariate data analysis*: Probability distributions and confidence intervals #. *Process monitoring*, aka statistical process control (SPC), for monitoring process behaviour. #. *Least squares regression modelling*: correlation, covariance, ordinary and multiple least squares models. Enrichment topics will be covered, time permitting. #. *Design and analysis of experimental data* and response surface methods for continual process improvement and optimization. #. Introduction to *latent variable modelling*: a general overview of latent variable models and their use in (chemical) engineering processes.

Several enrichment topics are covered throughout the course: robust methods, cross-validation for model assessment, nonparametric methods, real-time application of the above methods, correlation and causality, missing data handling, Bayesian methods, nonlinear regression.

Grading

=

To assess your understanding of the course materials, the grading for the course will be assessed as described below.

.. raw:: latex

\definecolor{orange}{rgb}{1,0.5,0}

{\color{orange}

Previous years (2010 to 2012)


.. tabularcolumns:: |p{3cm}|p{2.0cm}|p{10cm}|

.. csv-table::

  :header: "Component", "Fraction", "Notes"
  :widths: 15, 15, 40

"Assignments", "20%", "Expect around 7 assignments; can be completed individually, or in groups of 2 or less (4C3 and 6C3), or by yourself." "Midterm exam 1", "15%", "A 2.5 hour written exam, on 16 February, before the midterm break." "Midterm exam 2", "20 = 10+10%", "A take-home exam (10%), using software, over a 5-day period. An experimental report (10%) that you have to do with your group, ahead of time, and analyze the data from." "Final exam", "45%", "A written exam, lasting 3 hours."

.. raw:: latex

}

New outline for 2013 and future years


.. tabularcolumns:: |p{3cm}|p{2.5cm}|p{10cm}|

.. csv-table::

  :header: "Component", "Fraction", "Notes"
  :widths: 15, 15, 40

"Assignments", "20%", "Expect around 6 or 7 assignments; can be completed individually (mandatory for 6C3), or in groups of 2 or less (4C3)." "Weekly tests", "22%", "Short weekly tests covering the material taught the previous week(s)." "Experimental report", "10%", "An experimental report that you have to do with your group and analyze the data from. Due electronically on 26 March." "Final exam", "48%", "A written exam, lasting 3 hours."

6C3 students will have extra questions on all assignments, tests and exams.

.. rubric:: Policies regarding the course and regarding grading

  • *Very important note*: **Achieving a grade of below 50% in the final exam will automatically imply failure in the course, with a grade of F, no matter what your other grades in tests and assignments are.**
  • Weekly tests: the two midterms that would normally be scheduled for this course (15% + 10%) have been removed and replaced by 10 to 12 smaller tests, totalling 22%. These tests may be answered from any device with internet access. These tests are *completed individually*. There will be multiple choice, short answer and long answer questions on the tests, with questions randomly generated per student. Many questions will be automatically graded and the results returned to you after completion of the test window. Some questions will require you to upload electronic files and images.

The testing window is a period of time from 18:00 on Sunday night until 03:00 Tuesday morning (a 33-hour window), however once you start the test on your device you will only have an hour to complete it. The intention is that you study the material taught in class the week prior and then take the test. Questions will occasionally be from work covered several weeks prior to the test. Solutions will be revealed at the end of the window and usually discussed in class on Tuesday.

The reason for using short, frequent tests is that there is ample evidence in the learning literature that the `testing effect <http://en.wikipedia.org/wiki/Testing_effect>`_ and that `spaced repetition <http://en.wikipedia.org/wiki/Spaced_repetition>`_ improve your retention and understanding of the material. This is especially true in 4C3/6C3 where the course structure has been carefully planned to be cumulative. Click on the links for some basic background reading.

The other advantage of smaller tests are that they they lower the stakes and reduce the pressure on you, for example if you are having a bad week, it will only affect a small test of around 2%, and not 10 to 15% of your overall course grade.

**Note**: *computerized weekly testing is an experimental feature of 4C3/6C3 in 2013. The instructor reserves the right to revert back to written midterm exams at any time during the course, giving proportional credit for any weekly tests that have already been written*. The reversion, should it occur, will have a weighting similar to previous years.

  • If a midterm test is scheduled, it will *be optional* and there is no make-up for it. If you choose not to write the midterm, or cannot write it due to illness or other reasons, then the usual approach will be followed: the percentage contribution from the midterm will be added to the final examination weighting.
  • The weekly tests and assignments will also test your ability to use computer software to help complete the questions. The 4th year chemical engineering lab has the course software installed in the event that you do not have access to a computer.
  • We strongly encourage you to complete the assignments in groups of no more than 2 members. The 6C3 students **must** complete assignments on their own.
  • You, and your group, will receive the greatest benefit if you each do **all** the questions yourselves. Arrange to meet and review your solutions, discussing various approaches.
  • Assemble a **single submission** for the group - the TA's will not grade loose sheets handed in after the first submission. All group submissions must clearly show the names of the group members.
  • You are defeating the purpose of the group-based assignment if you simply divide the assignment into sections, one for each group member. This is definitely not recommended, because you are losing out on the learning opportunity of seeing your mistakes and the group member's mistakes, and learning from them.
  • No sharing of any work may be done between groups for assignments. This includes handwritten documents and electronic files of any type. This will be strictly enforced. Please ensure that you have read the University’s academic integrity policy (part of which is reproduced below).
  • This is a large class of about 90 students, so late hand-ins interfere with the TA's ability to efficiently grade your assignments. Late assignments will be penalized by deducting **30%** per day for every late day. A grade of zero will be given for submissions handed in after the solutions are posted (usually within 2 days of assignment hand-in).
  • Emergencies and such arise, so each person has 2 "late day" credits for assignments. So you can hand in one assignment 2 days late, or 2 assignments each one day late, without penalty.
  • Grading of all work in this course will include contributions for clarity and organization of presentation. This is an important professional skill that you have now successfully developed since second year.
  • No make-ups will be given for missed assignments. No make-ups will be given for missed weekly tests.
  • Any paper-based materials (textbooks, notes, *etc*) are allowed during tests and exams. Electronic textbooks and resources are, unfortunately, not permitted at the final exam, but may be used at any other time during the course, including weekly tests.
  • Any calculator may be used during the tests and exams.
  • All assignments will be graded, and the mean of the best :math:`N-1` assignments used to calculate the assignment grade. You should expect :math:`N \approx 7`, and the assignments will be frequent at the start of the course, slowing down at the end.
  • The final percentage grades will be converted to letter grades using the Registrar's recommended procedure.
  • Adjustment to the final grades may be done at the discretion of the instructor.
  • The final exam will be cumulative, based on the entire semester's material.

Important notes

====
    • Class participation**: Please bring a calculator to every class.
    • Course software**

Use of a computer is required in the course. The R-language (http://www.r-project.org/) will be used, and is a freely available software package that runs on Linux, Apple and Windows computers. The software is available in the 4th year Chemical Engineering computer labs. Minitab (you can rent a 6-month version very cheaply), MATLAB, or Python may be used as well; you should not use Microsoft Excel. Where time permits, the TA and the instructor will post solutions in these languages. More details are posted on the course website.

    • Out-of-class access**

The instructor has `office hours posted on his website <http://learnche.mcmaster.ca/contact-info/>`_, or other times may be arranged by email.

The TA's for this course can be contacted by email - please see their addresses above. Try to send email from your McMaster account - email from personal accounts are sometimes discarded by spam filters.

    • Disclaimer**: The above outline **may be modified**, as circumstances change.

Academic integrity

=======

You are expected to exhibit honesty and use ethical behaviour in all aspects of the learning process. Academic credentials you earn are rooted in principles of honesty and academic integrity.

Academic dishonesty is to knowingly act or fail to act in a way that results or could result in unearned academic credit or advantage. This behaviour can result in serious consequences, e.g. the grade of zero on an assignment, loss of credit with a notation on the transcript (notation reads: “Grade of F assigned for academic dishonesty”), and/or suspension or expulsion from the university.

It is your responsibility to understand what constitutes academic dishonesty. For information on the various types of academic dishonesty please refer to the Academic Integrity Policy, located at http://www.mcmaster.ca/academicintegrity

The following illustrates only three forms of academic dishonesty:

  1. . Plagiarism, e.g. the submission of work that is not one’s own or for which other credit has been obtained.
  2. . Improper collaboration in group work: this point is particularly important and will be strongly penalized in this course.
  3. . Copying or using unauthorized aids in tests and examinations.


Important dates

====

A list of *tentative* dates is below. Some changes will occur as the course progresses. Please check the course website at least 3 times per week for updates:

.. tabularcolumns:: |p{4cm}|p{10cm}|

.. csv-table::

  :header: "Date", "Description"

"8 January 2013", "Overview class: review of course content and administrative issues" "9 January", "\1. *Data visualization* section starts" "15 January", "Assignment 1 due" "15 January", "\2. *Univariate data analysis* section starts" "25 January", "Assignment 2 due" "25 January", "\3. *Process monitoring* section starts" "01 February", "\4. *Least squares modelling* section starts" "01 February", "Assignment 3 due" "15 February", "Assignment 4 due" "15 February", "**Potential written midterm** (see notes above)" "18-24 February", "Midterm break" "26 February", "\5. *Design and analysis of experimental data* section starts" "1 March", "Assignment 5 due" "15 March", "Assignment 6 due" "22 March", "*Kipling*" "26 March", "**Experimental project due**" "01 April", "Assignment 7 due (might be cancelled)" "02 April", "\6. *Latent variable methods introduction* starts" "10 April", "Optional review class" "12 April", "Exams start"

</rst>