== book == Hands-On Machine Learning with Scikit-Learn and TensorFlow https://news.ycombinator.com/item?id=17772211 A high bias low-variance introduction to Machine Learning for physicists (bu.edu) == software == First step: installing python, jupyter, etc. I have several python versions on hardmaple, from apple, anaconda, and homebrew. I continue to need to tweak python versions and installation setups - a pain. For this project I'm using the one in /usr/local/opt/python3, python 3.6.5 (It used to be that the regular python didn't come with numpy etc and that pip or whichever package manager hit errors when installing all that. So for a time I was using conda as my default. Now it seems that virtualenv and pip and the regular tools - which ones are "standard" seems to be somewhat controversial, see for example https://towardsdatascience.com/which-python-package-manager-should-you-use-d0fd0789a250 - will install the numerical pstuff.) $ cd /usr/local/opt/python3 $ ./pip3 install --upgrade pip $ ./pip3 install virtualenv $ which virtualenv /usr/local/bin/virtualenv $ virtualenv --version $ virtualenv --help (installed with /usr/local/opt/python/bin/python3.6) $ cd ~/sugar/academics/term/2018-09-fall/ $ virtualenv env $ . env/bin/activate # activate virutual python (env) $ which pip /Users/mahoney/sugar/academics/term/2018-09-fall/machine_learning/env/bin/pip (env) $ which python /Users/mahoney/sugar/academics/term/2018-09-fall/machine_learning/env/bin/python Python 3.6.5 (env) $ pip install jupyter matplotlib numpy panda scipy scikit-learn ... Note that after this, "python", "python3", "python3.6" are all the same: $ ls -al ./env/bin/py* -rwxr-xr-x 1 mahoney staff 286 Jun 1 17:05 ./env/bin/pygmentize lrwxr-xr-x 1 mahoney staff 9 Jun 1 17:04 ./env/bin/python -> python3.6 -rwxr-xr-x 1 mahoney staff 2387 Jun 1 17:04 ./env/bin/python-config lrwxr-xr-x 1 mahoney staff 9 Jun 1 17:04 ./env/bin/python3 -> python3.6 -rwxr-xr-x 1 mahoney staff 13068 Jun 1 17:04 ./env/bin/python3.6 $ pip --version pip 10.0.1 ... (python 3.6) $ pip3 --version pip 10.0.1 ... (python 3.6) Also "pip" and "pip3" are the same The "Hands-on with Machine Learning ..." code and examples are in a github repo which I've clone here : $ git clone https://github.com/ageron/handson-ml.git And then to launch the notebooks : $ jupyter notebook & == data == https://en.wikipedia.org/wiki/Data_set#Classic_data_sets https://en.wikipedia.org/wiki/Iris_flower_data_set == other == https://blog.ycombinator.com/learning-math-for-machine-learning/ https://news.ycombinator.com/item?id=17664084