Final Project Comp 598

<<<<<<< HEAD

LaTeX

https://www.sharelatex.com/project/5650c66739c542e95bab4fd9

TODO

• Implement naive learners such as SVM, Random Forest, etc so we'll be able to compare them with our crazy DL algo
• Implement One Class SVM
• document feature pre-processing
• Implement RNN, CNN, and maybe DBN if we have the time
• Prepare slides for Wednesday?
• Feel free to add or modify

Features and Background Information

• https://github.com/AlexPiche/Seizure-Prediction/blob/master/Articles/Howbert2014.pdf
• https://github.com/MichaelHills/seizure-prediction

PCA and ICA

• https://github.com/AlexPiche/Seizure-Prediction/blob/master/Articles/Varatharajah.pdf

CNN

• https://irakorshunova.github.io/2014/11/27/seizures/
• https://github.com/AlexPiche/Seizure-Prediction/blob/master/Articles/Mirowski2009.pdf

RNN

• python testRNN.py
  • can modify the type of architecture if you import RNN from RNN2layers or RNN3layers
  • can modify the number of hidden units in the files RNN2layers and RNN3layers.

Tutorial

• https://github.com/craffel/Lasagne-tutorial
• http://www.wildml.com/2015/09/recurrent-neural-networks-tutorial-part-1-introduction-to-rnns/
• https://gist.github.com/justiceamoh/abe5efeff1fa8277febb

DBN

• https://github.com/AlexPiche/Seizure-Prediction/blob/master/Articles/Wulsin2010.pdf
• https://github.com/AlexPiche/Seizure-Prediction/blob/master/Articles/Turner2014.pdf

f03e90fab094236f8e22cac479164c6a7362442f

https://www.sharelatex.com/project/5650c66739c542e95bab4fd9

Hill's GitHub

https://github.com/MichaelHills/seizure-prediction