Computational Statistics
Spring semester 2017
General information
Lecturer  Martin Mächler 

Assistants  Niklas Pfister, Dominik Rothenhäusler 
Lectures  Thu 1315
HG
E 5
>>
Fri 0910 HG E 1.2 >> 
Exercises  Fri 1012 HG E 1.2 >> 
Course catalogue data  >> 
Literature 
T. Hastie, R. Tibshirani, J. Friedman. The Elements of Statistical Learning. Springer
J. E. Gentle. Elements of Computational Statistics. Springer W. N. Venables, B. D. Ripley. Modern Applied Statistics with S. Springer 
Exam and lecture attestation (Testat) 
In order to obtain ECTScredit points you
have to pass the written exam during the examination session. This
exam will include tasks to be solved on a computer with R. No course
attendance confirmation is required to subscribe for the exam. For
ECTS credits, you do not have to hand in your solution to the
exercises during the semester. Doctoral students who need (part of the) credit points, but who do not require a grade should talk to Martin Mächler at the beginning of the semester. In that case you will need to solve and submit at least 70% of the exercises. 
Course content
Abstract"Computational Statistics" deals with modern methods of data analysis (aka "data science") for prediction and inference. An overview of existing methodology is provided and also by the exercises, the student is taught to choose among possible models and about their algorithms and to validate them using graphical methods and simulation based approaches.
ObjectiveGetting to know modern methods of data analysis for prediction and inference. Learn to choose among possible models and about their algorithms. Validate them using graphical methods and simulation based approaches.
Course Synopsismultiple regression, nonparametric methods for regression and classification (kernel estimates, smoothing splines, regression and classification trees, additive models, projection pursuit, neural nets, ridging and the lasso, boosting). Problems of interpretation, reliable prediction and the curse of dimensionality are dealt with using resampling, bootstrap and cross validation.
NoticeExercises will be based on the opensource statistics software R. Emphasis will be put on applied problems. Active participation in the exercises is strongly recommended.
Announcements

First week
Beginning of lecture: Thursday, 23/02/2017.Exercises start on 24/02/2017 at 10:15 in room HG E 1.2, with a special introduction to the "R" software. Please bring a laptop. The tutorial can be downloaded here. Here are some further helpful links related to R:
R homepage
R studio homepage
Getting help with R
Online R course (in German)
Try R

No lecture on Friday May 5th
There will be no lecture on Friday 05/05/2017, however the exercise will take place as usual. 
Question hours:
Monday, 07/08/2017, 2pm  3pm, HG G19.1
Thursday, 10/08/2017, 2pm  3pm, HG G19.1 
Exam Review:
Friday, 22/09/2017, 12pm  1pm, HG G19.1
Course materials
Lecture notes
The lecture notes are available here.
Recorded lectures
The recorded lectures can be found here. You need nethz login credentials to log in.
R Scripts as used in the lecture
A selection is online in this directory.
Course organisation
The course outline can be found here.
Exercise classes
Exercise classes will be held weekly on Friday from 10.15 to 11:55 in HG E 1.2. For the precise dates see the table below.
The exercise class on Friday will be divided in 2 parts. During the first part, the assistant will discuss the new exercise series. Part of the material covered during the lecture can also be recapitulated if this is requested. During the second part, the assistant will answer individual questions (regarding the lecture, the exercise sheet or R) and give you back the corrected solutions you submitted the previous exercise class.
Questions and comments
You can always ask questions to the lecturer and tutors during the class and the exercise session. If you have concrete R questions, you can bring your own laptop to the exercise class and ask them directly to the assistant. If you want to send us an email, please send it to this address: compstat@stat.math.ethz.ch. Please do not send any R code to this address, it is usually easier to answer these questions during the exercise class or to arrange an appointment.
Exercise sheets
The new exercise sheet will always be uploaded on the Thursdays preceding the corresponding preliminary discussion session.
We will correct your answers given that you respect our handin policy:

No R script files and no lengthy compilations of
outputs or figures.
 Only hand in your most important findings and answers. Do not include your code unless the question specifically asks to fill in a given skeleton of an R code.

Focus on the interpretation of the results.
 We are usually more interested in what you conclude based on the obtained results, than on the numbers per se.

Hand in by 12pm (noon) on due date.
 The solved exercises should be handed in during the exercise class or placed in the corresponding tray in HG J68 on the due date by 12pm at the latest.
A very elegant way to hand in your solution is to combine everything in a single file (for example by generating a pdf with LATEX or Word/Libreoffice). An easy way to do this is to use RMarkdown! See the tutorial and the template.
Exercises  Discussion  Deadline 

RTutorial  February 24, 2017  
Exercise 1  March 3, 2017  March 10, 2017 
Exercise 2  March 10, 2017  March 17, 2017 
Exercise 3  March 17, 2017  March 24, 2017 
Exercise 4  March 24, 2017  March 31, 2017 
Exercise 5, RSkeleton_Ex_5.2  March 31, 2017  April 7, 2017 
Exercise 6  April 7, 2017  April 28, 2017 
Exercise 7, skeleton7.R  April 28, 2017  May 5, 2017 
Exercise 8, skeleton8.R  May 5, 2017  May 12, 2017 
Exercise 9  May 12, 2017  May 19, 2017 
Exercise 10, skeleton10.R  May 19, 2017  May 26, 2017 
Exercise 11 (minor changes in 1. d); June 1st)  May 26, 2017  June 2, 2017 
General discussion  June 2, 2017 
Solutions
The solutions will be sent weekly to the students enrolled for this course.
Help with R
On Friday, February 24st 2017, there will be an introduction to the statistical software R during the exercise class from 10:15 to 11:55. The tutorial can be downloaded here. Here are some further helpful links related to R:
R homepage
R studio homepage
Getting help with R
Online R course (in German)
Try R