alburke / ams-2020-ml-python-course

Materials for the Machine Learning in Python for Environmental Science Problems AMS 2020 Short Course

ams-2020-ml-python-course

Machine Learning in Python for Environmental Science Problems AMS 2020 Short Course

Authors

• Amanda Burke, University of Oklahoma ([email protected])
• Benjamin Toms, Colorado State University ([email protected])
• Katherine Avery, University of Oklahoma ([email protected])
• Hamid Kamangir, Texas A&M Corpus Christi ([email protected])
• Karthik Kashinath, Lawrence Berkeley National Laboratory ([email protected])
• Ryan Lagerquist, University of Oklahoma ([email protected])

Modules

Introduction to Machine Learning

1. Introduction to Machine Learning and AI
2. Data Science Fundamentals
3. Supervised Learning Algorithms
4. Introduction to Deep Learning

Advanced Topics in Machine Learning

1. Unsupervised Learning Overview
2. Machine Learning Model Interpretation

Requirements

The modules for this short course require Python 3.6 and the following Python libraries:

• numpy
• scipy
• matplotlib
• xarray
• netcdf4
• pandas
• scikit-learn
• tensorflow-gpu or tensorflow
• keras
• jupyter
• ipython
• jupyterlab
• ipywidgets

Data Access

The data for the course are stored online. The download_data.py script will download the data to the appropriate location and extract all files. The netCDF data is contained in a 2GB tar file, so make sure you have at least 4GB of storage available and a fast internet connection.

Course Website

To run the notebooks on the cloud rather than a local installation, see the short course website Machine Learning in Python for Environmental Science.

ams-2020-ml-python-course's Issues

Comments for @hkaman7 's notebook

@hkaman7 I realize you might not be completely done with your notebook, but I wanted to give some feedback and ideas while we have time to look at it and adjust things.

• I think you're lecture 3, not lecture 2.
• Add a references section
• I crashed on the import utils line. This isn't really an issue; you can just have everyone add to the PYTHONPATH. I personally like adding a little boilerplate at the top to avoid changing environment variables manually:

import sys
from os.path import dirname, abspath
sys.path.insert(1, abspath('')+"/introduction_to_machine_learning")
...
import Lecture_2.utils as utils

• In the "Read Data" section, I already explore the contents of one of the files in my lecture.
• Amanda talks about normalization in her section, and I touch upon it briefly in relation to content in my section. Review what's in those sections and adjust content from your lecture as necessary.
• I go over linear regression, hyperparameter search, and L1 and L2 regularization in my section. I don't go over ElasticNet.
• At the moment, I introduce the idea of binarization, ROC curves, attribute diagrams, and performance diagrams at the end of my presentation (as we discussed earlier). Feel free to look at my stuff to see how you want to transition. I can delete/alter content at the end of my lecture depending on what you need for yours.
• Perhaps you could briefly introduce artificial neural nets to set up for Karthik's deep learning talk? (Very basic ANNs aren't necessarily deep learning, so I feel like that would go in your section fine.)
• I would explain SVMs as a concept, even if you don't have time to do a full-fledged experiment with one.