Event_Description.mkd

Introduction to Machine Learning / python / scikit-learn workshop

The idea is to help people get started with machine learning. It will be interactive and we'll be using scikit-learn (scikit-learn.org) for the machine Learning part and matplotlib (matplotlib.org) for visualizations.

This is the first time I'm doing this. It's gonna be (w)hacky. I won't cover much (read, any) theory. We'll use some publicly available datasets and play around with them.

Prerequisites:

• Laptop
• Do the following on your laptop
  • Download / install Virtualbox from https://www.virtualbox.org/wiki/Downloads
  • Download the NeuroDebian Virtual Machine here http://neuro.debian.net/debian/vm/NeuroDebian_7.2.0_amd64.ova
  • Open Virtualbox
  • Follow http://neuro.debian.net/vm.html to import/create the VM
• Or use your own scikit-learn / matplotlib installation
• Some python if you want. Plenty of tutorials online, google found this: http://www.learnpython.org
• Ability to look at a command prompt without nervous breakdown.

Topics are gonna be some of the following,

• Basics of Machine Learning
  • Classification
  • Regression
• Dimensionality Reduction
  • Principal Component Analysis
  • Implementation
• Visualization
• Linear Discriminant Analysis
• Support Vector Machines
  • Grid Search
• Ensembles
  • Random Forests
• ...

Cost: free for DSL members, HKD 200 otherwise. Goes pretty much entirely towards rent and utilities of DSL.