The Machine Learning “Advent Calendar” Day 20: Gradient Boosted Linear Regression in Excel

Gradient Boosting is an ensemble learning method that builds a model sequentially by training new models to correct the errors of the previous ones. This process is demonstrated by applying it to a linear regression base model, showing the step-by-step calculations in a spreadsheet. The article explains that while boosting linear regression is a purely pedagogical exercise with no practical advantage, it effectively illustrates the core mechanics of the algorithm. This iterative process of adding new models is framed as a form of Gradient Descent operating in a function space.