Experimental Physics (spring 2011)

Guest lecture/session on exploratory data analysis in MATLAB.

Course description
Absalon

Learning goals

The student is expected to recognise the following concepts and apply the methods to relevant datasets. No theoretical understanding is expected at this point.

• Multivariate analysis
  • Covariance Matrix
  • Correlations
  • Autocorrelations (optional)
• Principal Component Analysis
  • Linear models
  • Non-linear models (kernel methods)
• Visualisation
• Fast fourier transforms - cool blog post (optional)

Exercises

• Slides (PDF)
• Exercises in multivariate methods
• Exercise 0
• Covariances and Correlations (Exercise 1 & 2)
• Solutions for Exercise 1 & 2
• Principal Component Analysis (Exercise 3)
• Solutions for Exercise 3

Projects

• Project: Cosmic Rays
• Project: Cosmic Rays (PCA)
• Project: R-Hadrons at the LHC (Correlations)
• Project: Measuring velocity by doppler shift

Extra material

• Python iPod CSV to MATLAB script

On a linux or mac computer, the netcat command-line tool can be used for online logging of iSeismometer data:

nc -l -u 10552 > output.txt

Bibliography

Principal Component Analysis - An introduction - Thomas B. Moeslund

Singular Value Decomposition and Principal Component Analysis - Michael E. Wall, Andreas Rechtsteiner Luis M. Rocha