iw381/mvcomp2

Computational Statistics and Data Analysis (MVComp2 exercise class)

Master's Course, SS2015 Faculty of Physics and Astronomy, University of Heidelberg

Computational Statistics and Data Analysis (MVComp2)

Course LSF entry

Lecturer: PD Dr. Coryn Bailer-Jones

Assistants: Dr. Morgan Fouesneau, Dr. Dae-Won Kim

Summer semester 2015

This page provides the homework assignments in a form of python notebooks. You can click in the table below to read it online or download them. (python is not imposed to solve the exercises.)

This course and exercises take a pragmatic approach to using statistics and computational methods to analyse data. The focus will be on concepts, understanding problems, and the application of techniques to solving problems, rather than reproducing proofs or teaching you recipes to memorize.

The course website is available here

This repository gives the homeworks related datasets and eventually corrections (table below for links)

The repository will be updated after each class to give the assignments. All datasets, gists of code will also be included. Examples of solutions (hardly unique) will be included eventually.

Some homework guidelines

Notebooks have no meaning of imposing a format to give us back your homework assignments. Instead they give me convenient ways to keep both texts and codes at the same place.

• Each week, we will mark your homework on a scale of 100 points in total. (details given with the exercises)

• You are allowed to work in groups of at most 3 persons and return 1 document per group.

• Homework documents must be returned each Tuesday.

• We do not mark your coding skills.

• This means we do not read the codes. We do not look out for comments in the codes, but we will not guess what a plot means. Be explicit and describe even in once sentence what you did.

• Feel free to use the notebooks (it may not be the most efficient), be careful when printing (Check out nbconvert to produce a pdf or even latex document).

Computing language

• We do not impose a language. Feel free to use any that you judge efficient for you. Obviously we cannot provide full support, nor we cannot give full tutorials.

• If you use R, many examples of code will be included in the lecture notes. If you use Python, all the exercises will be using python (when coding is required).

• examples in R from the course are available here: link (will be updated throughout the course)

Online tools

In case you cannot/do not want to install libraries or softwares on your computer, some free online services exist, such as:

Sage Cloud: python, R, and other languages

Wakari Python only.

MCMC libraries

some libraries that you may find useful later depending on your language.

emcee (python)

STAN (C/C++)

Lectures

There will be 12 lectures on the following dates (the exercise session is on the following day). The topics allocated to the dates may well change!

As github now integrates nbviewer If a notebook is not accessible through the links in the table, you can instead click on the files

Lecture date	Topic	Exercises	datasets & snippets
14 April	Introduction and probability basics	notebook	rvs.dat
21 April	Estimation and error: describing data and distributions	notebook	star.csv
28 April	Statistical models and inference	notebook	hipparcos.dat
5 May	Linear models and regression	notebook	rmr_ISwR.dat sdss_sspp_sub.csv
12 May	(Bayesian) Model fitting I	notebook	lighthouse.dat
19 May	(Bayesian) Model fitting II	notebook	coinflip.dat
26 May	MCMC	notebook	2Dline.dat metropolis.py
2 June	No lecture
9 June	Hypothesis testing	notebook	iswr_vitcap.dat
16 June	Model Comparison	notebook	2Dline_modelcomparison.dat
23 June	Cross validation, regularization, and basis functions	notebook	ratdiet_fields.dat cars93sel_MASS.dat
30 June	Kernels and other things
7 July	Classification
14 July	Study week
23 July	Exam