Articles
-
Must we adjust p-values if we test multiple hypotheses?
May 16, 2024
tl; dr Yes, yes you must
-
Optimal Bayesian Sequential Hypothesis Testing
February 2, 2024
tl; dr We introduce the mSPRT and give a derivation from a Bayesian point of view.
-
“What p-values really mean“
January 28, 2024
tl;dr Most people get p-values wrong. This is how to understand and apply them correctly.
-
An introduction to Gradient Descent
October 10, 2021
tl; dr Gradient Descent is the simplest learning algorithm. It is very easy to implement, is robust to complex cost functions, and is the foundation of an enormous zoo of variations.
-
Time Series with GAMs
August 13, 2021
tl; dr Generalised Additive Models are a flexible class of models which improve on polynomial regression. This post shows how to fit time series data with GAMs a bit like
prophet. -
Structural Time Series in PyMC!
August 13, 2021
tl; dr Bayesian Structural Time Series let us model time series component by component. Here’s how to do it in PyMC
-
The guts of GAMs
August 10, 2021
tl; dr Generalised Additive Models look harder than they actually are. We fit a single dimension GAM from scratch, including generating splines.
-
Neural Networks from scratch.
June 21, 2021
tl; dr Neural Networks are just mathematical expressions. In this post we create a small NN library, including autograd, to show how the internals work.
-
Bayesian Structural Time Series in pystan.
June 7, 2021
tl; dr Bayesian Structural Time Series let us model time series component by component. Here’s how to do it in Stan
-
More Complex (Linear) Regressions
March 5, 2021
tl; dr We extend our ideas of how to do regression to a more complex class of functions.
-
Local Linear Trend models for time series
March 2, 2021
tl; dr Local Linear Trend models are one of the simplest time series models. Here we code them up in python.
-
Mixed Models
January 1, 2021
tl; dr Group Structure is something we can exploit to improve our estimates. We introduce and explain mixed models and how to fit them.