PHYS 453 - Spring 2023, Dr. Mike Daugherity, Abilene Christian University.
- Tutorial 1 - Python Cheat Sheet: Reminders of logic, loops, and formatting output
- Tutorial 2 - Numpy: Doing math with arrays in numpy
- Tutorial 3 - matplotlib: Making plots with matplotlib
- Class Notes: Summary from class 1.23.2023
- Cheat Sheets: numpy | matplotlib | pandas | sklearn
- PPT 0 - First Day slides: A too-short intro to data science and machine learning, and why it is taught by a physicist
- Tutorial 4 - First_Look_at_Data: Using scikit-learn and pandas to look at real data
- PPT 1 - Intro To Classifiers: The basic vocabulary of supervised learning and classifiers
- Homework 1 - DIY_1D_Classifier: Make your own simple classifier! Make that machine learn!
- Tutorial 5 - A_Classifier_Class: Lets give our classifier an interface like the scikit-learn classifiers
- Homework 2 - Won't You Be My Neighbor: Make a simple Nearest Neighbor classifier
- PPT 2 - Evaluating Classifiers: Notes on how to score a classifier. With Math!
- Classifier Challenge: Evaluating classifiers problem in class 2.22.23
- Tutorial 6 - Tuning and Evaluation: How to tune and evaluate performance of a classifier
- Tutorial 7 - Data Transforms: The unskippable step of scaling data
For most classifiers I have Powerpoint slides explaining how it works and tutorial code for how to use it.
- Nearest Neighbor: PPT 3 | Tutorial 8
- Decision Trees: PPT 4 | Tutorial 9
- Tutorial 10 - Titanic Pandas: Using pandas to look at the Titanic dataset
- Homework 3 - Trees on the Titanic: Find out who sinks and who swims on real Titanic data
- Class Challenge: Spring Break is over, let's remember how to do machine learning
- Bayesian Classifier: PPT 5 | Tutorial 11
- Tutorial 12 - Digits Dataset: Exploring the digits dataset for HW4
- Homework 4 - Digits Throwdown: Show off each classifier's special tricks on the digits dataset
- Linear Models: PPT 6
- Neural Networks: PPT 7 | Tutorial 13
Copy-and-paste code to get you started on a problem
- Classification Recipe: Code to get you started on a classification problem using pipelines and grid searches
- Regression Recipe: regression problems using pipelines and grid searches (doesn't include discussion on cleaning data)
- PPT 8 - Transforms - feature scaling, dimensionality reduction with PCA, manifold learning and t-SNE, pipelines
- PPT 9 - Regression - predicting a real-valued number instead of a category
- PPT 10 - Unsupervised Learning - what we can learn without labeled training data: clustering, genetic evolution, random (stochastic) methods
- PPT 11 - Ensembles - improving performance with multiple classifiers/regressors: bagging, adaboost, random forests, gradient boosted trees, XGBoost
- PPT 12 - Modern Methods - what's new in machine learning: GANS, Stable Diffusion, Deep Learning, LLMs