Dear enrolled or wait-listed student, ECS 172, Winter 2022,

I have a few points that you may find important:

1.  We have a long waiting list for the course.  We will try to get a
larger room, to expand the course enrollment limit.  However, this
cannot be done until mid-December.

2.  Several people have asked me if it is OK to miss the class'
discussion section due to time conflicts.  Unfortunately, the discussion
section is REQUIRED.  The exams, ie the weekly quizzes, will be given in
the discussion section.

3.  One person asked if students can use Python instead of R in the
course.  Unfortunately, the answer here again is no.  You will be using
R recommender systems libraries, will do the assignments in groups of 3
or 4 students in R, and will take the quizzes in R.  BTW, you may be
interested in my comparison of Python and R for DATA SCIENCE
APPLICATIONS, https://github.com/matloff/R-vs.-Python-for-Data-Science.

Below, I'm enclosing the course blurb that I had sent out previously.

Norm


*********************************************************************
*********************************************************************

New Course:  ECS 172, Recommender Systems

Winter 2022

Prof. Norm Matloff 

Days/Time:  Lecture MWF, 9-9:50; discussion W 10-10:50

Prerequisites:  

   * Linear algebra: MAT 22A or equivalent.  Matrix algebra, rank,
   dimension, etc.

   * Calculus-based probability:  ECS 132 or equivalent (e.g. STA 131A
   or MAT 135A).  Density functions, expected value, etc.

   * Programming:  We will use R, but background in any language is fine.

Further comments on background:

   Linear algebra and probability will be reviewed at the beginning of
   the course, using recommender systems datasets as examples.

   We will use machine learning methods but this is NOT a prerequisite.
   The material will be developed in the course, again using recommender
   systems as examples/applications.

   Non-CS majors welcome!

Description:

   Collaborative filtering and content-based methods for building
   recommender systems.  Machine learning, statistical, matrix
   factorization and nearest-neighbor approaches.  Use of sentiment
   analysis.  Google PageRank. Case studies.

Workload:

   * Comparable to typical non-programming upper-division CS courses,
   say ECS 152A.  (Less intensive than my ECS 132 class. :-) )

   * Weekly or biweekly open-book quizzes (no midterm or final exam);
   group assignments (mainly analysis of real data); group term project.  

The field of recommender systems (RS):

   * A branch of machine learning. 

   * Prediction of how well a given "user" will like a given "product."
   Amazon, Spotify, OK Cupid, etc.

   * The University of Minnesota developed a course adviser system.  A
   student uses it to predict how well she would like a certain
   course.  It even predicts her grade!

Cool video:  

   See the great Stanford presentation by Pooja Rajkumar, then a UCD undergrad, https://www.youtube.com/watch?v=G_z4sXiYGog, on RS software she and other UCD students developed.