This is it! Final wrap-up of STAT 750
What is a statistical model
- An abstraction, a summary.
- Deterministic equation + likelihood function
(mathematical formula of the (linear) predictor + probability distribution)
- Steps to design a statistical model:
- Define the distribution the data arise from.
- Define the linear predictor.
- Select a method to estimate the parameters.
- Least Squares
- Maximum Likelihood Estimation
- REML
- Bayesian Estimation (check of STAT 786 - Applied Bayesian Modeling and Prediction)
Key things to report
- Expected value.
- Uncertainty!!!
Statistics as a catalyzer of scientific progress
- The connection between scientific questions and statistics.
Miscellaneous
- Next class: review linear regression, multiple linear regression.
- Sources where to learn how to code better without relying on AI: