IEOR (산업공학)과 학생들이 과 밖에서 들은 수업에 대해 개인적인 견해를 적은 글을 올립니다.
일단, 모든 의견들은 개인적 견해일 뿐이니 그저 참고만 하세요.
혹시 맘에 안 드는 글귀나 의견이 있으셔도 너그러이 보아주시기 바랍니다. 반대 의견을 올려 주셔도 좋구요
어디까지나 비공식적인, 그냥 서로 의견 교환하자고 만든 파일이니까요.
Information on some of the classes offered outside the
department
December, 2003
The following information was provided by some of our students who have taken classes in
other departments. It is intended to help IEOR students in finding useful courses outside
the department. Note, that in some cases different students had different opinion about
the same course.
The following few paragraphs refer to several STAT classes:
Most people from other departments do the 200A and B when they minor in Statistics. I
did 200B in spring and am currently doing the 200A. Good courses, decent and thorough
introduction to many useful statistical techniques and also the core theoretical stu?.
The 205 and 210 are really, really hard with a lot of Lebesgue integration, measure theory
etc. Some people with a lot of math from ECON take these courses for their minor, but
be warned.
STAT 101 (Introduction to the Theory of Probability)
Prof. Klass, Spring 2002; Hoel, Port and Stone
If you haven't had probability this is it, introductory probability theory for Stat majors
(134 is for engineers and considered weaker). The book (Hoel, Port and Stone) is terse
and pithy, but very good, and if you do not know the material already you need to read
it because Klass' lectures leave out a lot of "need to know information." One either loves
Klass or hates him (more than half the class disappeared after the midterm). His style
is extemporaneous, disorganized and interactive. His overhead manners are atrocious,
but he has a severe disability which interferes with his handwriting. He proposes various
problems cold and in solving them often dips into advanced mathematics beyond the class
level. He offered some neat probability solving tricks I've never seen before or since and
really encouraged thinking deeply about new problems. I thought he was a softy when it
came to grading. On a scale from 1 to 10, it gets an 8 from me (though I would guess the
average lower).
STAT 155 (Game Theory)
² Erol Pekoz, Fall 1997
A fun and easy undergrad course which introduces a lot of topics in game theory.
There are a lot of homeworks just as any other undergrad course but none of them
are hard. The instructor is not at Berkeley any more.
² Undergrad course for game theory. The instructor changes every semester, thus
hard to expect same thing will be o?ered, but very interesting.
STAT 200B (Intro to Statistics)
David Brillinger, Spring 2003; Charles Stone, A Course in Probability and Statistics"
This course was certainly not the most challenging I have taken. It was however a not-too-painful introduction to the language of statistics. If you have always wondered what a P-value and null hypothesis were and will never force yourself to figure it out independently,then this may be a good class for you. There were two inclass midterms and a final, and all were open book though the final was supposed to be closed book according to the syllabus. There were also labs, which were approximately biweekly exercises in fiddling with data and using some of the statistics we'd been talking about. This course assumes
a little knowledge in probability, and I believe Stat 200A, which is an introduction to
probability, uses the same textbook. It isn't the best textbook in the world, but there
will be regular homeworks out of it, so if you are taking the class you probably want to
go ahead and get your own copy.
STAT 205A (Probability Theory)
² J. Pitman, Fall 1998
The course to introduce probability in a much more theoretical way. It is recom-
mended for students who are interested in stochastic processes. Weekly homework
and a take home final determined the grade. As I have sat in the lecture by another
instructor, personally I think Pitman is the one to go with.
² J. Pitman, Fall 1995, R. Durrett Probability Theory"
Covers basic measure theory. Recomended to anyone who is intending to concentrate
on probabilistic modelling. Very work intense and requires solid knowledge of real
analysis.
STAT 215 (Statistical Models: Theory and Applications)
² The 215 is a very good course. But 4 units is a joke, more like 6 or 7. Very hard,
labs every week, many people learn new ways of cursing through the semester. But
eventually, you learn how to apply Splus to many situations. I would, though,
have at least taken 200A first. How would you be able to know how to simulate
distributions by uniform variables, calculate transformations of random variables,
joint distributions and stu? otherwise?
² David Brillinger, Fall 1996
There was no assigned text.
I took the course from Brillinger in Fall 1996. We had 9 labs a midterm and a
final. The labs are quite labor-intensive and must be typed up in latex. The course
assumes that students posses good understanding of theoretical statistics before
entry.
The course was ok, but there may be better choices in the stat department. The
professor is nice and approachable, but his lectures are somewhat disorganized.
This class is good but a lot of work (it earns its 4 units). You'll learn a lot about
linear models and extensions. The class uses Splus (a statistical software package)
heavily. Be prepared to spend a lot of time on the computers. IEOR computers
also have Splus, so you're not bound to use the stat computers, though it is nice to
have that option. Their grad lab is pretty nice, but their quotas are tiny.
² The lecture is a lot of theory (linear algebra, and some optimization), which doesn't
always immediately connect to the assignments. Have a good background in sim-
ple stats (especially hypothesis testing) before starting this class. Also, the class
assumes you have some intuition about both theoretical and practical matters in
stats that stat majors might have, but you'll have to develop as you go.
² David Brillinger
Prof. Brillinger is good, nice to talk to, and interested in any outside projects you
bring.
The usual classroom, 1011 Evans, has a nice view of the bay.
STAT 215B Statistical Models: Theory and Application (4 units)
Course Format: Three hours of lecture and two hours of laboratory per week.
The techniques of applied statistics. Data types and structures. Model formulation, ¯t-
ting and validation. The principal models. Planning and design. Difficulties that arise.
Usage of statistical computer packages. Presentation of conclusions. (F,SP)
(From the 1999-2001 General Catalog updated as of 4/21/99)
² Prof. Speed, Spring 1996, 2 readers
This course had two readers (about 4 inches thick in total). The above description of
the course is somewhat misleading. In this course we read several statistical studies
a week and then critiqued them. There were no exams that I recall. The grade
came largely from class participation and on a few written assignments. I took the
course from Speed in Spring 1996. I would not recommend this course.
² Prof. Freedman
I took it w/ Prof. Freedman. He is a very sharp guy and a good teacher, but the
class was kinda too hard for a non-stat major. It resembles IEOR 254 in the sense
that the whole class is based on papers that you read and then discuss in the lecture.
Then there were labs, homeworks, a term project and a ¯nal, too. He used not to
go through the math details of the papers but it was impossible to discuss them
properly if you did not do so. In the end the most useful experience I got from that
course was learning LaTeX, Splus and critisizing papers and see how people tend
to obscure facts when they write papers. It covered a wide variety of topics, linear
models in particular. Later in some instances I had the opportunity to use them
and I found the course pretty useful. But if this course is taught by a di?erent prof
then I'd expect it to be pretty different.
STAT 230 (Linear Models)
² 230 is a very useful course. Many econ students take that. 232 and 236 also. 242
is said to be good. 248 is supposed to be extremely useful in econ (time series), but
very hard.
² Prof. Bullman, Fall 1995; Seber, Linear Regression"
(Don't buy the textbook if very expensive.) Covers various linear regression models
and uses S-plus for lab. exercises. Very useful if want to learn about linear regres-
sion. Known to be more theoretical when taught by other instructors but is a very
good course.
STAT 241A (Statistical Learning Theory); also CS 281A
² Michael Jordan, Fall 2002; reader
A very good introduction to probabilistic/graphical methods for statistical inference
problems. well motiviated, and covered in detail. definitely recommend it. Jordan
is a great instructor and the homeworks are good. time consuming, but you de¯-
nitely get a good grasp of the material by doing them. I don't think you need any
background for this class beyond basic probability and calculus. that said, some
graph theory/linear algebra and makes the class much easier. the homeworks were
a good amount of work, but well worth it.
² Michael Jordan, Fall 2002; reader
I greatly enjoyed this course and found him to be a good lecturer, and fairly good
at answering questions that students posed. Because there were students from a
variety of different departments in this class, the questions were also varied, and
at many different levels. The homework and final project were challenging and
time consuming. Many of the homeworks included some sort of coding work, which
could apparently be done in C or C++, but I chose to work with Matlab. I think it
would be important to be able to program algorithms in Matlab to do this course,
otherwise the coding parts of the homework take up much more time than they are
worth. Also I think a good understanding of matrix manipulation would be useful.
The discussion sections taught by David Blei were also very useful for getting more
fundamental material and asking the questions that you thought were too silly to
ask in class.
The course was fairly fast-paced and we covered a lot of material. If you think you
may be working on parameter estimation, this would probably be a good class for
you, especially since you can work on pretty much whatever you want for the ¯nal
project, so you can do some research and get graded on it as well.
² Michael Jordan, Fall 2000; reader
(The reader we had was a draft of a book he was working on on this topic. Perhaps
it has been published by now.) I liked this course - the graphical approach to prob-
ability is quite nice, and a LOT easier than the algebra one is typically confronted
with in trying to calculate conditional probabilities, etc. When I took it, there were
no exams, the grade was 80% homeworks and 20% ¯nal project. It is a challenging
course, and at times I felt Jordan assumed too much statistical knowledge (e.g. de-
tailed properties of the multinomial Normal distribution). However, he is happy to
explain things when asked. You will be extremely familiar with Matlab by the time
the course is over. I recommend it.
² Michael Jordan, Fall 2000; reader
I took it with Michael Jordan, who's an EXCELLENT lecturer and can really keep
you interested during the lectures. The material is very interesting: it gives a bisic
probabilistic theory and computational methods for the statistical modeling of com-
plex, multivariate data. The emphasis is on the connection between the probability
theory and graphical models.
Prerequisites: The prerequisites for this course include previous coursework in lin-
ear algebra, multivariate calculus, and basic probability and statistics. Previous
coursework in graph theory, information theory, optimization theory and statistical
physics would be helpful but is not required. Students will need to be familiar with
Matlab, Splus or a related matrix-oriented programming language.
Couple of other good things about this course: it is listed in both the Stat and
CS departments, so you can use it towards the minor in either. Also, there's no
textbook to buy: you get the notes from Jordan's upcoming book: An Introduction
to Probabilistic Graphical Models, for free.
STAT 243 and 244 (Statistical Computing)
² 243 is easy, 244 hard. I don't know how easy it is to skip 243, though.
² Phil Spector
Excellent course if you like programming and want to learn C. The ¯rst six weeks
the prof (Phil Spector) just covers the basics of C and then you start learning
stat algorithms. The topics covered were mainly mean and variance computations,
random number generation, matrix operations, linear and nonlinear regression. I
loved the course (mainly b/c I love programming) and I think Spector is a very nice
prof and also a good teacher. It may be a little hard for somebody who has not
done programming before, though. There are only 4 assignment in the course, but
14
they tend to take a lot of time.
Note: 243 and 244 are only taught by Phil Spector.
² Phil Spector, Fall 1999
Good introduction to C and Splus with useful algorithms, mostly statistical (linear
and nonlinear regression) but matrix manipulations, Newton method are relevant
to IEOR. 4 programming homeworks in a semester.
STAT 248 (Time Series Analysis)
² David Brillinger, Spring 2003; Brockwell and Davis, Introduction to Time Series
and Forecasting"
This is the same text as is used in the undergraduate version of this course, and for
the ¯rst 8 weeks or so, I felt as if I was in the undergrad course. Then the course got
crazy. Brillinger is a funny and nice person, but is probably not the best lecturer,
and this came out the strongest when we moved on to more complicated material,
including spectral stu?, a Huge Part of most aspects of time series. Luckily the
book wasn't too awful and when it came time to do the ¯nal project, which is
a huge portion of the grade, I was able to put something together by following
the book. It was interesting, and not the most time consuming class o?ered at
the university, but not the easiest either. However, my experiences may not be
representative of what you will ¯nd if you take this class, because Dr. Brillinger
frequently admitted to drastically changing the course structure and content from
year to year to keep from being bored. If you are Canadian or like hockey, you will
have a strong extracurricular bond with him.
² David Brillinger, Spring 2003; Brockwell and Davis
I think that you get what you want out of the course, after about half way through
the semester things were just too high level for me to profoundly understand them
but that's ok because I wasn't trying to.
Probably if you're into ¯nance this would be a very interesting class. Also the
professor is quie entertaining and if anything you learn about the wide application
of time series analysis
² David Brillinger, Spring 2002; no required texts
(No reqd textbook, but i suggest you get the recommended text.) Brillinger is very
funny in class. however, to really learn the stuff, you will have to work quite a bit
outside of the class. there are no exams, the homeworks are trivial, and the project
can really be anything even vaguely related with time series. so what you make of
the class boils down to how much extra work you do.
² David Brillinger, Spring 2002; no required texts
This course gives a good intro to time series. It is definitely not a strenuous" course,
15
though it can be a little frustrating at times as the lectures seem disorganized. Doing
the labs at the time they are given probably would help with understanding, as would
more relevant homeworks. On the whole, though, I think it is a course worth taking
- things will come together when you're working on your final project at the very
latest. Brillinger is a great guy, and very interested in different" project ideas.
² David Brillinger, Spring 2000
This course will be most useful if you already have a data in your research and want
to carry out time series analysis on it. Very few homeworks and 2 projects.
STAT 260 (Topics on Statistics, Seminar)
Prof. Doksum, Fall 1995; text: Fan, Gibbons ????
Content varies depending on instructors interest. Doksum covers non-parametric, semi-
parametric models and topics that are of students interest. Only requires a ¯nal project
and HW's, but only recomended to those who are interested in complex statistial models.
STAT 261 (Quantitative/Statistical Research Methods in Social
Sciences)
Prof. Goodman
261 is a waste of time, Goodman is not a good teacher.
Engineering (various)
Note: I was unable to ¯nd some of these courses in the general catalog and/or the
department websites. In some cases, the courses are listed, but with different names.
==========================
Civil and Environmental Engineering
Transportation
CIV ENG 193 (Engineering Risk Analysis)
Prof. Masoud Zadeh, Fall 2003
This class will not be particularly challenging if you have taken Ross's 263 course. To-
ward the end of the course, we covered some interesting material on parameter estimation
techniques. Beyond that, I thought the most beneficial part of the course was the appli-
cations. They tried discuss the applications of each method we learned in the context in
which it would be used in the real world.
Civil Engineering
CIV ENG 251 (Operation of Transportation Facilities)
Prof. Cassidy, Fall 2002; Daganzo Fundamentals of Transportation and Tra±c Opera-
tions"
Traffic flow theory is the main topic of this course. Many useful tools are discussed and
the approach is fairly non-theoretical. These tools are not limited to the area of trans-
portation, but can be used in many different fields. I'm happy to have taken this course
and I would recommend it to other IEOR grads. The workload is normal.
CIV ENG 254 (Transportation Economics)
Prof. Hansen, Spring 2002?
Good things: No Midterm, only 5 (or is it 4) Hw for the semester and 1 big project(which
counts to 25% of grade) and take home exam. NO TEXT BOOKS, everything is available
online. Other good things,you got to learn how to use new software TSP which are quite
common in Economterics. Oh and the professor is quite funny. One more thing, it counts
to Logistics certificate. Bad Things: Tough Homeworks..it might not be many, but it
takes quite a lot of time to do (especially if you're not familiar with Excel, Economics, or
TSP software). I consider the course a quite advanced economics course and it involves a
lot of econometrics..so if you are not familiar, be ready to work hard.
CIV ENG 258 (Logistics)
² Prof. Daganzo, Fall 2002; Daganzo, Logistic Systems Analysis"
This course is a more advanced and theoretical course whose main focus is on dis-
tribution systems. The professor is very relaxed and informal. A big part of the
course consists of a semester-long project (instead of a ¯nal exam), which was very
interesting. The goal of our project was to model and optimize the Chinese postal
system, a not so simple task. It was a very rewarding course to take, I felt that I
learned a lot. I would recommend it to other IEOR grads. The workload is normal.
² Prof. Daganzo, Fall 2002; Daganzo, Logistic Systems Analysis"
Great class, I understand much more about calculating transporataion cost and de-
sign methdology of logistic system. The term project is espcially good re-enforcement
of materials learned in class.
CIV ENG 260 (Air Transportation)
Prof. Hansen
We covered some mechanics of airplanes, operational things such as airport capacity
planning, noise abatement, crew and plane scheduling, etc... I would recommend it as a
practical course.
CIV ENG 260L (Air Transportation, Lab)
Kanafani
We planned an airport expansion for St. Louis airport. Included runway capacity analysis.
I would recommend it - again, a very practical course.
CIV ENG 263 (Operations of Transportation Terminals)
² Prof. Daganzo, Spring 2003; no required text
This is really good class, Prof Daganzo brings great insights on some of the applica-
tion of IEOR modeling tools. The course work isn't too heavy either, students have
the option to do a term project or turn in 6 out of 10 homeworks. So the students
can decide how much time and e?ort they want to invest in that class.
² Prof. Daganzo
We covered terminal operations including rail, marine, and tra±c issues. The tech-
niques were di?erent from traditional OR, but still useful as an alternate perspective.
Techniques allow for back-of-the-envelope calculations, and are especially useful
where only aggregate or incomplete data is available.
² Daganzo, Fall 1996, reader
Good practical tools that can be used for quick and dirty analysis and then can be
re¯ned with more rigorous formulae or simulation. There was no assigned text, but
we had a reader. I took the course from Daganzo in Fall 1996. I would de¯nitely
recommend the course with Daganzo. The course is completely assignment driven.
No exams as I recall.
CIV ENG 264A-C (Networks, Logistics)
Prof. Newell, Prof. Daganzo
A little esoteric, but good if transportation is your focus. Again, non-traditional opti-
mization techniques, but adds to one's background knowledge of "continuous modeling".
City and Regional Planning
CIV ENG 213 (Transportation and Land Use Planning)
Prof. Cervero
A little unsophisticated mathematically, but provided good insights to urban transit plan-
ning. Instructor gave very good examples from all over the world.
CIV ENG 214 (Infrastructure Planning)
Prof. Dowall
Instructor was very competent. Course material was a little o? the beaten path as far as
what IEOR students should be learning. (not really mathematical)
CIV ENG 40 (Urban Form)
Prof. Southworth
Interesting class, but again, topic was a little divergent from the content one might want
as an OR student. (not really mathematical)
Comment: In general, I thought all instructors were good. I would recommend all classes
subject to one's interest in the subject matter and focus of the minors (i.e., practical vs.
theoretical). As for texts, in almost all instances, course readers were used.
일단, 모든 의견들은 개인적 견해일 뿐이니 그저 참고만 하세요.
혹시 맘에 안 드는 글귀나 의견이 있으셔도 너그러이 보아주시기 바랍니다. 반대 의견을 올려 주셔도 좋구요
어디까지나 비공식적인, 그냥 서로 의견 교환하자고 만든 파일이니까요.
Information on some of the classes offered outside the
department
December, 2003
The following information was provided by some of our students who have taken classes in
other departments. It is intended to help IEOR students in finding useful courses outside
the department. Note, that in some cases different students had different opinion about
the same course.
The following few paragraphs refer to several STAT classes:
Most people from other departments do the 200A and B when they minor in Statistics. I
did 200B in spring and am currently doing the 200A. Good courses, decent and thorough
introduction to many useful statistical techniques and also the core theoretical stu?.
The 205 and 210 are really, really hard with a lot of Lebesgue integration, measure theory
etc. Some people with a lot of math from ECON take these courses for their minor, but
be warned.
STAT 101 (Introduction to the Theory of Probability)
Prof. Klass, Spring 2002; Hoel, Port and Stone
If you haven't had probability this is it, introductory probability theory for Stat majors
(134 is for engineers and considered weaker). The book (Hoel, Port and Stone) is terse
and pithy, but very good, and if you do not know the material already you need to read
it because Klass' lectures leave out a lot of "need to know information." One either loves
Klass or hates him (more than half the class disappeared after the midterm). His style
is extemporaneous, disorganized and interactive. His overhead manners are atrocious,
but he has a severe disability which interferes with his handwriting. He proposes various
problems cold and in solving them often dips into advanced mathematics beyond the class
level. He offered some neat probability solving tricks I've never seen before or since and
really encouraged thinking deeply about new problems. I thought he was a softy when it
came to grading. On a scale from 1 to 10, it gets an 8 from me (though I would guess the
average lower).
STAT 155 (Game Theory)
² Erol Pekoz, Fall 1997
A fun and easy undergrad course which introduces a lot of topics in game theory.
There are a lot of homeworks just as any other undergrad course but none of them
are hard. The instructor is not at Berkeley any more.
² Undergrad course for game theory. The instructor changes every semester, thus
hard to expect same thing will be o?ered, but very interesting.
STAT 200B (Intro to Statistics)
David Brillinger, Spring 2003; Charles Stone, A Course in Probability and Statistics"
This course was certainly not the most challenging I have taken. It was however a not-too-painful introduction to the language of statistics. If you have always wondered what a P-value and null hypothesis were and will never force yourself to figure it out independently,then this may be a good class for you. There were two inclass midterms and a final, and all were open book though the final was supposed to be closed book according to the syllabus. There were also labs, which were approximately biweekly exercises in fiddling with data and using some of the statistics we'd been talking about. This course assumes
a little knowledge in probability, and I believe Stat 200A, which is an introduction to
probability, uses the same textbook. It isn't the best textbook in the world, but there
will be regular homeworks out of it, so if you are taking the class you probably want to
go ahead and get your own copy.
STAT 205A (Probability Theory)
² J. Pitman, Fall 1998
The course to introduce probability in a much more theoretical way. It is recom-
mended for students who are interested in stochastic processes. Weekly homework
and a take home final determined the grade. As I have sat in the lecture by another
instructor, personally I think Pitman is the one to go with.
² J. Pitman, Fall 1995, R. Durrett Probability Theory"
Covers basic measure theory. Recomended to anyone who is intending to concentrate
on probabilistic modelling. Very work intense and requires solid knowledge of real
analysis.
STAT 215 (Statistical Models: Theory and Applications)
² The 215 is a very good course. But 4 units is a joke, more like 6 or 7. Very hard,
labs every week, many people learn new ways of cursing through the semester. But
eventually, you learn how to apply Splus to many situations. I would, though,
have at least taken 200A first. How would you be able to know how to simulate
distributions by uniform variables, calculate transformations of random variables,
joint distributions and stu? otherwise?
² David Brillinger, Fall 1996
There was no assigned text.
I took the course from Brillinger in Fall 1996. We had 9 labs a midterm and a
final. The labs are quite labor-intensive and must be typed up in latex. The course
assumes that students posses good understanding of theoretical statistics before
entry.
The course was ok, but there may be better choices in the stat department. The
professor is nice and approachable, but his lectures are somewhat disorganized.
This class is good but a lot of work (it earns its 4 units). You'll learn a lot about
linear models and extensions. The class uses Splus (a statistical software package)
heavily. Be prepared to spend a lot of time on the computers. IEOR computers
also have Splus, so you're not bound to use the stat computers, though it is nice to
have that option. Their grad lab is pretty nice, but their quotas are tiny.
² The lecture is a lot of theory (linear algebra, and some optimization), which doesn't
always immediately connect to the assignments. Have a good background in sim-
ple stats (especially hypothesis testing) before starting this class. Also, the class
assumes you have some intuition about both theoretical and practical matters in
stats that stat majors might have, but you'll have to develop as you go.
² David Brillinger
Prof. Brillinger is good, nice to talk to, and interested in any outside projects you
bring.
The usual classroom, 1011 Evans, has a nice view of the bay.
STAT 215B Statistical Models: Theory and Application (4 units)
Course Format: Three hours of lecture and two hours of laboratory per week.
The techniques of applied statistics. Data types and structures. Model formulation, ¯t-
ting and validation. The principal models. Planning and design. Difficulties that arise.
Usage of statistical computer packages. Presentation of conclusions. (F,SP)
(From the 1999-2001 General Catalog updated as of 4/21/99)
² Prof. Speed, Spring 1996, 2 readers
This course had two readers (about 4 inches thick in total). The above description of
the course is somewhat misleading. In this course we read several statistical studies
a week and then critiqued them. There were no exams that I recall. The grade
came largely from class participation and on a few written assignments. I took the
course from Speed in Spring 1996. I would not recommend this course.
² Prof. Freedman
I took it w/ Prof. Freedman. He is a very sharp guy and a good teacher, but the
class was kinda too hard for a non-stat major. It resembles IEOR 254 in the sense
that the whole class is based on papers that you read and then discuss in the lecture.
Then there were labs, homeworks, a term project and a ¯nal, too. He used not to
go through the math details of the papers but it was impossible to discuss them
properly if you did not do so. In the end the most useful experience I got from that
course was learning LaTeX, Splus and critisizing papers and see how people tend
to obscure facts when they write papers. It covered a wide variety of topics, linear
models in particular. Later in some instances I had the opportunity to use them
and I found the course pretty useful. But if this course is taught by a di?erent prof
then I'd expect it to be pretty different.
STAT 230 (Linear Models)
² 230 is a very useful course. Many econ students take that. 232 and 236 also. 242
is said to be good. 248 is supposed to be extremely useful in econ (time series), but
very hard.
² Prof. Bullman, Fall 1995; Seber, Linear Regression"
(Don't buy the textbook if very expensive.) Covers various linear regression models
and uses S-plus for lab. exercises. Very useful if want to learn about linear regres-
sion. Known to be more theoretical when taught by other instructors but is a very
good course.
STAT 241A (Statistical Learning Theory); also CS 281A
² Michael Jordan, Fall 2002; reader
A very good introduction to probabilistic/graphical methods for statistical inference
problems. well motiviated, and covered in detail. definitely recommend it. Jordan
is a great instructor and the homeworks are good. time consuming, but you de¯-
nitely get a good grasp of the material by doing them. I don't think you need any
background for this class beyond basic probability and calculus. that said, some
graph theory/linear algebra and makes the class much easier. the homeworks were
a good amount of work, but well worth it.
² Michael Jordan, Fall 2002; reader
I greatly enjoyed this course and found him to be a good lecturer, and fairly good
at answering questions that students posed. Because there were students from a
variety of different departments in this class, the questions were also varied, and
at many different levels. The homework and final project were challenging and
time consuming. Many of the homeworks included some sort of coding work, which
could apparently be done in C or C++, but I chose to work with Matlab. I think it
would be important to be able to program algorithms in Matlab to do this course,
otherwise the coding parts of the homework take up much more time than they are
worth. Also I think a good understanding of matrix manipulation would be useful.
The discussion sections taught by David Blei were also very useful for getting more
fundamental material and asking the questions that you thought were too silly to
ask in class.
The course was fairly fast-paced and we covered a lot of material. If you think you
may be working on parameter estimation, this would probably be a good class for
you, especially since you can work on pretty much whatever you want for the ¯nal
project, so you can do some research and get graded on it as well.
² Michael Jordan, Fall 2000; reader
(The reader we had was a draft of a book he was working on on this topic. Perhaps
it has been published by now.) I liked this course - the graphical approach to prob-
ability is quite nice, and a LOT easier than the algebra one is typically confronted
with in trying to calculate conditional probabilities, etc. When I took it, there were
no exams, the grade was 80% homeworks and 20% ¯nal project. It is a challenging
course, and at times I felt Jordan assumed too much statistical knowledge (e.g. de-
tailed properties of the multinomial Normal distribution). However, he is happy to
explain things when asked. You will be extremely familiar with Matlab by the time
the course is over. I recommend it.
² Michael Jordan, Fall 2000; reader
I took it with Michael Jordan, who's an EXCELLENT lecturer and can really keep
you interested during the lectures. The material is very interesting: it gives a bisic
probabilistic theory and computational methods for the statistical modeling of com-
plex, multivariate data. The emphasis is on the connection between the probability
theory and graphical models.
Prerequisites: The prerequisites for this course include previous coursework in lin-
ear algebra, multivariate calculus, and basic probability and statistics. Previous
coursework in graph theory, information theory, optimization theory and statistical
physics would be helpful but is not required. Students will need to be familiar with
Matlab, Splus or a related matrix-oriented programming language.
Couple of other good things about this course: it is listed in both the Stat and
CS departments, so you can use it towards the minor in either. Also, there's no
textbook to buy: you get the notes from Jordan's upcoming book: An Introduction
to Probabilistic Graphical Models, for free.
STAT 243 and 244 (Statistical Computing)
² 243 is easy, 244 hard. I don't know how easy it is to skip 243, though.
² Phil Spector
Excellent course if you like programming and want to learn C. The ¯rst six weeks
the prof (Phil Spector) just covers the basics of C and then you start learning
stat algorithms. The topics covered were mainly mean and variance computations,
random number generation, matrix operations, linear and nonlinear regression. I
loved the course (mainly b/c I love programming) and I think Spector is a very nice
prof and also a good teacher. It may be a little hard for somebody who has not
done programming before, though. There are only 4 assignment in the course, but
14
they tend to take a lot of time.
Note: 243 and 244 are only taught by Phil Spector.
² Phil Spector, Fall 1999
Good introduction to C and Splus with useful algorithms, mostly statistical (linear
and nonlinear regression) but matrix manipulations, Newton method are relevant
to IEOR. 4 programming homeworks in a semester.
STAT 248 (Time Series Analysis)
² David Brillinger, Spring 2003; Brockwell and Davis, Introduction to Time Series
and Forecasting"
This is the same text as is used in the undergraduate version of this course, and for
the ¯rst 8 weeks or so, I felt as if I was in the undergrad course. Then the course got
crazy. Brillinger is a funny and nice person, but is probably not the best lecturer,
and this came out the strongest when we moved on to more complicated material,
including spectral stu?, a Huge Part of most aspects of time series. Luckily the
book wasn't too awful and when it came time to do the ¯nal project, which is
a huge portion of the grade, I was able to put something together by following
the book. It was interesting, and not the most time consuming class o?ered at
the university, but not the easiest either. However, my experiences may not be
representative of what you will ¯nd if you take this class, because Dr. Brillinger
frequently admitted to drastically changing the course structure and content from
year to year to keep from being bored. If you are Canadian or like hockey, you will
have a strong extracurricular bond with him.
² David Brillinger, Spring 2003; Brockwell and Davis
I think that you get what you want out of the course, after about half way through
the semester things were just too high level for me to profoundly understand them
but that's ok because I wasn't trying to.
Probably if you're into ¯nance this would be a very interesting class. Also the
professor is quie entertaining and if anything you learn about the wide application
of time series analysis
² David Brillinger, Spring 2002; no required texts
(No reqd textbook, but i suggest you get the recommended text.) Brillinger is very
funny in class. however, to really learn the stuff, you will have to work quite a bit
outside of the class. there are no exams, the homeworks are trivial, and the project
can really be anything even vaguely related with time series. so what you make of
the class boils down to how much extra work you do.
² David Brillinger, Spring 2002; no required texts
This course gives a good intro to time series. It is definitely not a strenuous" course,
15
though it can be a little frustrating at times as the lectures seem disorganized. Doing
the labs at the time they are given probably would help with understanding, as would
more relevant homeworks. On the whole, though, I think it is a course worth taking
- things will come together when you're working on your final project at the very
latest. Brillinger is a great guy, and very interested in different" project ideas.
² David Brillinger, Spring 2000
This course will be most useful if you already have a data in your research and want
to carry out time series analysis on it. Very few homeworks and 2 projects.
STAT 260 (Topics on Statistics, Seminar)
Prof. Doksum, Fall 1995; text: Fan, Gibbons ????
Content varies depending on instructors interest. Doksum covers non-parametric, semi-
parametric models and topics that are of students interest. Only requires a ¯nal project
and HW's, but only recomended to those who are interested in complex statistial models.
STAT 261 (Quantitative/Statistical Research Methods in Social
Sciences)
Prof. Goodman
261 is a waste of time, Goodman is not a good teacher.
Engineering (various)
Note: I was unable to ¯nd some of these courses in the general catalog and/or the
department websites. In some cases, the courses are listed, but with different names.
==========================
Civil and Environmental Engineering
Transportation
CIV ENG 193 (Engineering Risk Analysis)
Prof. Masoud Zadeh, Fall 2003
This class will not be particularly challenging if you have taken Ross's 263 course. To-
ward the end of the course, we covered some interesting material on parameter estimation
techniques. Beyond that, I thought the most beneficial part of the course was the appli-
cations. They tried discuss the applications of each method we learned in the context in
which it would be used in the real world.
Civil Engineering
CIV ENG 251 (Operation of Transportation Facilities)
Prof. Cassidy, Fall 2002; Daganzo Fundamentals of Transportation and Tra±c Opera-
tions"
Traffic flow theory is the main topic of this course. Many useful tools are discussed and
the approach is fairly non-theoretical. These tools are not limited to the area of trans-
portation, but can be used in many different fields. I'm happy to have taken this course
and I would recommend it to other IEOR grads. The workload is normal.
CIV ENG 254 (Transportation Economics)
Prof. Hansen, Spring 2002?
Good things: No Midterm, only 5 (or is it 4) Hw for the semester and 1 big project(which
counts to 25% of grade) and take home exam. NO TEXT BOOKS, everything is available
online. Other good things,you got to learn how to use new software TSP which are quite
common in Economterics. Oh and the professor is quite funny. One more thing, it counts
to Logistics certificate. Bad Things: Tough Homeworks..it might not be many, but it
takes quite a lot of time to do (especially if you're not familiar with Excel, Economics, or
TSP software). I consider the course a quite advanced economics course and it involves a
lot of econometrics..so if you are not familiar, be ready to work hard.
CIV ENG 258 (Logistics)
² Prof. Daganzo, Fall 2002; Daganzo, Logistic Systems Analysis"
This course is a more advanced and theoretical course whose main focus is on dis-
tribution systems. The professor is very relaxed and informal. A big part of the
course consists of a semester-long project (instead of a ¯nal exam), which was very
interesting. The goal of our project was to model and optimize the Chinese postal
system, a not so simple task. It was a very rewarding course to take, I felt that I
learned a lot. I would recommend it to other IEOR grads. The workload is normal.
² Prof. Daganzo, Fall 2002; Daganzo, Logistic Systems Analysis"
Great class, I understand much more about calculating transporataion cost and de-
sign methdology of logistic system. The term project is espcially good re-enforcement
of materials learned in class.
CIV ENG 260 (Air Transportation)
Prof. Hansen
We covered some mechanics of airplanes, operational things such as airport capacity
planning, noise abatement, crew and plane scheduling, etc... I would recommend it as a
practical course.
CIV ENG 260L (Air Transportation, Lab)
Kanafani
We planned an airport expansion for St. Louis airport. Included runway capacity analysis.
I would recommend it - again, a very practical course.
CIV ENG 263 (Operations of Transportation Terminals)
² Prof. Daganzo, Spring 2003; no required text
This is really good class, Prof Daganzo brings great insights on some of the applica-
tion of IEOR modeling tools. The course work isn't too heavy either, students have
the option to do a term project or turn in 6 out of 10 homeworks. So the students
can decide how much time and e?ort they want to invest in that class.
² Prof. Daganzo
We covered terminal operations including rail, marine, and tra±c issues. The tech-
niques were di?erent from traditional OR, but still useful as an alternate perspective.
Techniques allow for back-of-the-envelope calculations, and are especially useful
where only aggregate or incomplete data is available.
² Daganzo, Fall 1996, reader
Good practical tools that can be used for quick and dirty analysis and then can be
re¯ned with more rigorous formulae or simulation. There was no assigned text, but
we had a reader. I took the course from Daganzo in Fall 1996. I would de¯nitely
recommend the course with Daganzo. The course is completely assignment driven.
No exams as I recall.
CIV ENG 264A-C (Networks, Logistics)
Prof. Newell, Prof. Daganzo
A little esoteric, but good if transportation is your focus. Again, non-traditional opti-
mization techniques, but adds to one's background knowledge of "continuous modeling".
City and Regional Planning
CIV ENG 213 (Transportation and Land Use Planning)
Prof. Cervero
A little unsophisticated mathematically, but provided good insights to urban transit plan-
ning. Instructor gave very good examples from all over the world.
CIV ENG 214 (Infrastructure Planning)
Prof. Dowall
Instructor was very competent. Course material was a little o? the beaten path as far as
what IEOR students should be learning. (not really mathematical)
CIV ENG 40 (Urban Form)
Prof. Southworth
Interesting class, but again, topic was a little divergent from the content one might want
as an OR student. (not really mathematical)
Comment: In general, I thought all instructors were good. I would recommend all classes
subject to one's interest in the subject matter and focus of the minors (i.e., practical vs.
theoretical). As for texts, in almost all instances, course readers were used.
