MODULE 1 - The Foundational Underpinnings of Machine Learning-Course Introduction
Course overview: Machine Learning Under the Hood
()
Why this course isn't hands-on & why it's essential for techies anyway
The Machine Learning Glossary
One-question survey
MODULE 1 - The Foundational Underpinnings of Machine Learning-Ensuring Discoveries Are Trustworthy
P-hacking: a treacherous pitfall
()
P-hacking: your predictive insights may be bogus
()
P-hacking: how to ensure sound discoveries
()
Complementary materials on p-hacking (optional)
Avoiding overfitting: the train/test split
()
MODULE 1 - The Foundational Underpinnings of Machine Learning-Correlation Does Not Imply Causation
Why ice cream is linked to shark attacks
()
Causation is just a hobby -- prediction is your job
()
Correlation does not imply causation (optional)
MODULE 1 - The Foundational Underpinnings of Machine Learning-The Principles of Predictive Modeling
The art of induction: why generalizing from data is hard
()
Learning from mistakes: why negative cases matter
()
Intro to the hands-on assessment (Excel or Google Sheets)
()
Data access for auditors (optional)
MODULE 2 - Standard, Go-To Machine Learning Methods-Decision Trees: a Great Place to Start
A refresher on decision trees
()
Business rules rock and decision trees rule
()
Pruning decision trees to avoid overfitting
()
DEMO - Comparing decision tree models (optional)
()
A powerful, helpful visualization of how decision trees work (optional)
Drawing the gains curve for a decision tree
()
Drawing the profit curve for a decision tree
()
MODULE 2 - Standard, Go-To Machine Learning Methods-Beyond Trees: Other Standard Modeling Methods
Naïve Bayes
()
Linear models and perceptrons
()
Linear part II: a perceptron in two dimensions
()
Why probabilities drive better decisions than yes/no outputs
()
Logistic regression
()
DEMO - Training a logistic regression model (optional)
()
MODULE 3 - Advanced Methods, Comparing Methods, & Modeling Software-Advanced Modeling Methods
How neural networks work
()
Neural nets: decision boundaries & a comparison to logistic regression
()
DEMO - Training a neural network model (optional)
()
Deep learning
()
Ensemble models and the Netflix Prize
()
Supercharging prediction: ensembles & the generalization paradox
()
DEMO - Training an ensemble model (optional)
()
DEMO - Autotuning a machine learning model (optional)
()
The generalization paradox of ensembles (optional)
MODULE 3 - Advanced Methods, Comparing Methods, & Modeling Software-Modeling Methods Overview: Summary, Software, and Deployment
Compare and contrast: summary of ML methods
()
Machine learning software: dos and don'ts for choosing a tool
()
Machine learning software: how tools vary and how to choose one
()
Model deployment: out of the software tool and into the field
()
MODULE 3 - Advanced Methods, Comparing Methods, & Modeling Software-Uplift Modeling (aka Persuasion Modeling)
Uplift modeling I: optimize for influence and persuade by the numbers
()
Uplift modeling II: modeling over treatment and control groups
()
Uplift modeling III: how it works – for banks and for Obama
()
Uplift modeling IV: improving churn modeling, plus other applications
()
Complementary readings on uplift modeling (optional)
MODULE 4 – Pitfalls, Bias, and Conclusions-Ethics: Machine Bias, Model Transparency, and Conclusions
Machine bias I: the conundrum of inequitable models
()
The original ProPublica article on machine bias
Machine bias II: visualizing why models are inequitable
()
Interactive MIT Technology Review article on disparate false positive rates
Another interactive demo of machine bias (optional)
Machine bias III: justice can't be colorblind
()
Complementary reading on machine bias (optional)
Explainable ML, model transparency, and the right to explanation
()
More on explainable ML and model transparency (optional)
Conclusions on ML ethics: establishing standards as a form of social activism
()
Tallying the positive and negative impacts of AI (optional)
MODULE 4 – Pitfalls, Bias, and Conclusions-Specialization Wrap-Up
Pitfalls: the seven deadly sins of machine learning
()
John Elder's top ten data science mistakes (optional)
Conclusions and what's next – continuing your learning
()
Further resources and readings to continue your learning (optional)