Machine Learning Lab (MLDM Grad Program)

Authors: Austin Schwinn & Usama Javaid

Dates: Jan - Mar 2017

Subject: Repository for my grad school machine learning lab

Lab 1:

Exercise 1: Scikit-Learn practice with Iris dataset. Matplotlib.
Exercise 2: Create Kayak Race CSV in data subfolder 
Exercise 3: KNN with Scikit & Iris
Exercise 4: Decision Trees with Scikit & Iris.
Exercise 5: Neural Networkds & Multilayer Perceptron Classifier for digit recognition
Exercise 6: Using original Kayak Race CSV to test KNN, Decision Trees, and MLP NN

Note: Tree output visualizations available in Outputs subdirectory of Data and Dependencies

Lab 2:

Exercise 1: Hidden Markov Model (HMM) with weather example
Exercise 2: HMM with string and states
Exercise 3: HMM for sequence of string states

Lab 3:

Exercise 1: SVM simple example
Exercise 2: SVM with randomly generated data set
Exercise 3: SVM with Moon dataset
Exercise 4: SVM with Iris dataset
Exercise 5: SVM with Ozone dataset

Note: Ozone dataset is provided by MeteoFrance. Below are description for the attributes.
	- JOUR: type of day, holiday (1) or not (0).
	- O3obs: ozone concentration observed the next day at 17h (local time), often at the maximum pollution rate.
	- MOCAGE: prediction of the pollution made by a deterministic model
	- TEMPE: temperature for the next day at 17h.
	- RMH2O: humidity rate
	- NO2: nitrogen dioxide concentration
	- NO: nitric oxide concentration
	- STATION: place from where the observations are taken (Aix-en-Provence, Rambouillet, Munchhausen, Cadarache or Plan de Cuques)
	- VentMOD: wind strength
	- VentANG: wind orientation

References:

Based on lab in Machine Learning course of Machine Learning and Data Mining (MLDM) Master's Program and University Jean Monnet. Course taught by Dr. Elisa Fromont and Dr. Amaury Habrard.

MLDM Program Webpage: http://mldm.univ-st-etienne.fr/

MeteoFrance & Ozone Dataset