ML-DA

Repository for "Machine Learning and Data Analytics" course @ university of Trieste, A.Y. 2017/2018

Datasets usually are either provided along with the python jupyter notebooks, or they are taken from the repository ferdas's faraway repository, which is a collections of datasets converted from the faraway R package.

"First time only" config

Python3 is required: if you installed python 3 using your distribution package manager or by compiling it from source you should already have pip3 installed, otherwise install it following these instructions.

I suggest to use a virtual-environment (virtualenv) to set up a dedicated sandbox for this project. Moreover, to better manage different virtualenvs, I suggest to use virtualenvwrapper.

These are the instructions to follow:

1. install virtualenvwrapper for every user on your pc, with root permissions:

# pip3 install virtualenvwrapper

2. add the needed config to your .profile or .bash_profile file:

VIRTUALENVWRAPPER_PYTHON=/usr/bin/python3
export WORKON_HOME=$HOME/.venvs
source $(which virtualenvwrapper.sh)

3. then reload your .profile or .bash_profile file:

$ source ~/.bashrc

4. and eventually create a virtualenv for this project, using python3 as the python executable:

$ mkvirtualenv -p $(which python3) ML_DA

Normal usage

For a daily usage, activate the virtualenv created before:

$ workon ML_DA

Now you can manage it as a normal virtualenv.

To install the requirements use pip (and a virtualenv isolated sandbox is suggested to avoid messing up different packages from different projects), paste:

$ pip3 install -r python_requirements.txt

Enable the jupyter notebook

In the repository base folder just type:

$ jupyter-notebook