University of California at Berkeley
Dept. of Statistics

STAT210A
Theoretical Statistics

Fall Semester 2006



Practical information

See also the UCB On-Line Course Catalog and Schedule of Classes

Volume:

Lectures: Evans Hall 332, Tues--Thurs 9:30am--11am

Grading: Homeworks (15%), Midterm (25%) and Final Exam (60%).

Instructor: Martin Wainwright

Office Hours: Tues. 11am--12pm, Thurs. 1--2pm (421 Evans Hall)
Email: wainwrig AT stat DOT berkeley DOT edu
Phone: 643-1978
Offices: 263 Cory Hall or 421 Evans Hall

GSI: Guilherme V. Rocha

Office Hours: Mon. 5pm--6pm, Wed. 5pm--6pm (344 Evans),
Sessions : Thu 5pm--6pm (344 Evans)
Email: gvrocha AT stat DOT berkeley DOT edu
Office: 435 Evans Hall

Back to top

Course description

Overview: This course provides an introduction to mathematical statistics, including both frequentist and Bayesian aspects of modeling, prediction, inferences, hypothesis testing and estimation, as well as large sample properties and asymptotics. It presumes a solid background in undergraduate probability, real analysis, linear algebra as well as some degree of mathematical maturity.
Intended audience:

Required background: Undergraduate probability, linear algebra, real analysis; some degree of mathematical maturity.

Back to top

Updates and Announcements

  • Oct. 20: Solutions to the midterm can now be downloaded (file removed).

  • Oct. 14: Midterm will include material up to and HW6 and, from the lectures on the last week (10/09 to 10/13), only the concept of maximum likelihood estimation (defintion and how to compute it).

  • Oct. 14: Solutions to HW6 posted.

  • Oct. 10: Solutions to HW4 posted. Scribing notes for lecture 11 posted.

  • Oct. 09: Solutions to HW3 posted.

  • Oct. 04: Scribing notes for lectures 9 and 10 posted. Solutions to HW2 posted.

  • Sep. 21: Check material on exponential families and variational methods in the other useful material section below;

  • Sep. 18: Corrections in problem 3.3: Z is distributed according to a normal with mean \theta and variance \sigma^2 AND G(x)=P(Z<=x|Z>0), ie G is the conditional distribution of Z given that it is greater than 0.

  • Sep. 12: Change in Guilherme's OHs: New times are Mon 5-6PM and Thursdays 6-7PM. New section times: Thursdays 5-6PM at 332 Evans;

  • Sep. 12: Solutions to HW1 posted.

  • Sep. 07: Homework 2 assigned.

  • Sep. 06: Course mail list up and running: send an email to majordomo@listlink.berkeley.edu with empty subject and "subscribe stat210a" in the body of the message; 5B

  • Sep. 06: Homework 1: There is a typo in problem 1.2.b: the weighting factor should be the inverse of the variance \sigma_{i}^{2} and not the inverse of the standard deviation as posted. (G. Rocha)

  • Sep. 05: Homework 1: There is a problem in proving one of the implications in problem 1.5.b. The proof that sup \beta(\theta', \delta) >= \beta(\theta, \delta) for all theta'\in\Theta_1 and \theta \in \Theta_0 implies unbiasedness will NOT be graded.

  • Aug. 31: Homework 1 assigned, Scribing materials posted

    Back to top

    Lecture scribe notes

  • Scribe notes, Lecture 28, 12/07 (pdf)
  • Scribe notes, Lecture 27, 12/05 (pdf)
  • Scribe notes, Lecture 26, 11/30 (pdf)
  • Scribe notes, Lecture 25, 11/28 (pdf)
  • No scribe notes on 11/23: Thanksgiving
  • Scribe notes, Lecture 24, 11/21 (pdf)
  • Scribe notes, Lecture 23, 11/16 (pdf)
  • Scribe notes, Lecture 22, 11/14 (pdf)
  • Scribe notes, Lecture 21, 11/09 (pdf)
  • Scribe notes, Lecture 20, 11/07 (pdf)
  • Scribe notes, Lecture 19, 11/02 (pdf)
  • Scribe notes, Lecture 18, 10/31 (pdf)
  • Scribe notes, Lecture 17, 10/26 (pdf)
  • Scribe notes, Lecture 16, 10/24 (pdf)
  • No scribe notes on 10/19: Midterm examination
  • Scribe notes, Lecture 15, 10/17 (pdf)
  • Scribe notes, Lecture 14, 10/12 (pdf)
  • Scribe notes, Lecture 13, 10/10 (pdf)
  • Scribe notes, Lecture 12, 10/05 (pdf)
  • Scribe notes, Lecture 11, 10/03 (pdf)
  • Scribe notes, Lecture 10, 09/28 (pdf)
  • Scribe notes, Lecture 09, 09/26 (pdf)
  • Scribe notes, Lecture 08, 09/21 (pdf)
  • Scribe notes, Lecture 07, 09/19 (pdf)
  • Scribe notes, Lecture 06, 09/14 (pdf)
  • Scribe notes, Lecture 05, 09/12 (pdf)
  • Scribe notes, Lecture 04, 09/07 (pdf)
  • Scribe notes, Lecture 03, 09/05 (pdf)
  • Scribe notes, Lecture 02, 08/31 (pdf)
  • Scribe notes, Lecture 01, 08/29 (pdf)

    Back to top

    Handouts

  • Lecture 1 (Tues Aug 29): Course information (pdf) (ps)

    Back to top

    Instructions for homeworks

    Problem sets will be posted on the class webpage, and will be due on Thursdays in class at the start of lecture. Late homeworks will not be accepted. If they chose, after attempting the problems on an individual basis, students can discuss homework assignments in groups of at most three. However, each student must write up his/her own solutions individually, and must explicitly name any collaborators at the top of the homework.

    Homework Assignments (Solutions have been removed)

  • Homework 11 (Due Dec 07 in class) (file removed) Solutions: (file removed)

  • Homework 10 (Due Nov 16 in class) (file removed) Solutions: (file removed)

  • Homework 9 (Due Nov 9 in class) (file removed) Solutions: (file removed)

  • Homework 8 (Due Nov 2 in class) (file removed) Solutions: (file removed)

  • Homework 7 (Due Oct 26 in class) (file removed) Solutions: (file removed)

  • Homework 6 (Due Oct 12 in class) (file removed) Solutions: (file removed)

  • Homework 5 (Due Oct 5 in class) (file removed) Solutions: (file removed)

  • Homework 4 (Postponed: Due 4PM 09/29 in 435 or 367 Evans) (file removed) Solutions: (file removed)

  • Homework 3 (Due Sept 21 in class) (file removed) Solutions: (file removed)

  • Homework 2 (Due Sept 14 in class) (file removed) Solutions: (file removed)

  • Homework 1 (Due Sept 7 in class) (file removed) Solutions: (file removed)

    Back to top

    Scribing materials and instructions

    Lecture scribe notes should be written up in LATEX using the template and style files that can be downloaded from this page. Please send your written notes in PS or PDF format, as well as the original LATEX source, by email to Guilherme Rocha with the subject heading Lecture Scribe #N (where N is the lecture number), no later than 4 days after the lecture. The notes will be reviewed, and necessary editing/corrections will be indicated.

    Back to top

    Useful references

  • P. Bickel and K. Doksum, Mathematical Statistics.
    Available online (e.g., Amazon, Barnes and Noble).

  • R. Keener Statistical Theory: A Medley of Core Topics.
    Available as a reader at North Side Copy (Euclid & Hearst).

    • Additional References:
  • E.L. Lehmann and G. Casella. Theory of Point Estimation.
  • E.L. Lehmann and J. Romana. Testing Statistical Hypotheses.
  • Mark J. Schervish. Theory of Statistics.

    • For background on linear algebra:
  • Introduction to linear algebra, G. Strang.

    Other useful materials:
  • Graphical Models, exponential families and variational inference