Outline

Intro

LM

Core ideas
Prediction
Basic Interpretation
Adding Complexity
Assumptions of a linear (and other) model(s)

Question- how to fit classification here without spilling into other stuff?

Model Criticism

General fit
Model Comparison
Model Selection
Model debugging

Misc:

Model transparency (e.g. model cards)
Model fairness

Model Estimation

OLS, MLE
Classification
Penalized
Optimization, SGD
Bayesian

LM Extensions

GLM, GAM, Mixed Models
Optim/Linear Programming/Hungarian Algorithm?
Latent Linear Models
Mixture models/Clustering

Machine learning

CV, metrics
Lasso, Ridge, Elastic Net
Trees
- RF
- GBM
DL
- NN
- Autoencoders
Reinforcement learning
Ensemble models

Uncertainty

Bayesian inference
Bootstrap
Conformal Predictions

Data Stuff

Feature and Target Transformations
missing data
- recommend assessment of predicted data similarity to observed data
data quality and reliability/measurement
Sparsity
Outliers
Imbalanced data
'Big' Data, Scalability
Data types
- Categorical
- Ordinal
- Continuous
- Time series
- Text
- Images
- Audio
- Video
- Geospatial
- etc.
Feature Engineering/Pre-processing/Categorical Embeddings/Dimensionality Reduction/Feature Selection/Feature Extraction
misc feature types: ordinal, zero-infated, etc.
Transformations: std, log, max
Data leakage
Data drift
Data bias (lack of representativeness), vs. statistical bias
Misc:
- Data privacy, security, ethics
- Data provenance, governance

Other Stuff

Causality
- Causal inference
- Techniques: experimental design, matching, meta-learners, uplift modeling, etc.