Data Analysis for LEL - Week 9

Where to next?

Stefano Coretta

University of Edinburgh

Where to next?

Statistical modelling

A statistical model is a (simplified) mathematical representation of a process.

Linear models

Linear models are a type of statistical model.
In their simplest form, they are equivalent to the formula of a line.

\[ y = a + b \cdot x \]

Obscure terminology

See One Thousand and One names.

Statistical inference

Frequentist inference

Null Hypothesis Significance Testing (NHST).
Inference based on p-values (they don’t mean what most people think they mean).
MUST READ: Mindless statistics by Gigerenzer.

Bayesian inference

Inference based on the probabilities of possible values.
Re-use prior knowledge to make research a cumulative enterprise.
MUST READ: How to become a Bayesian in eight easy steps by Etz et al.

Machine learning

Beware the golems! See The Golem of Prague by McElreath.
Machine Learning is a mindset. See [Modeling Mindsets] by Molnar.
No interest in the how and why, just in predictive accuracy.

(Not) the end