Machine-learning notebooks

Jupyter notebooks on machine-learning algorithms. These are supplementary material for the AS4501 "Astroinformatics" and EL4106 "Computational Intelligence" courses at Universidad de Chile.

1. Neural networks: pure-numpy multilayer perceptron (MLP), tensorflow MLP and Bayesian MLP with PyMC3
2. Support vector machines: C-SVM, nu-SVM and one-class SVM using scikit-learn
3. Boosting with decision trees: Decision trees, Adaboost and Gradient boosting using scikit-learn
4. Bagging with decision trees: Decision trees, Bagging and Random Forest using scikit-learn
5. Self organizing maps: Color clustering through SOM using Somoclu

Requirements will vary between notebooks. Incomplete list of dependecies:

• Python 3 (not tested with Python 2)
• Numpy
• Tensorflow
• PyMC3
• Theano
• Scikit-learn
• Somoclu