How to streamline your data with Tensorflow — Imagine that you have a dataset (X, y) containing a huge amount of features that you want to use to solve a task such as regression or classification. While having more observations is always better, including too many features into a model can be counterproductive, be it for interpretability or…

Make my mediocre marketing mix model much more mighty — Flashback In my last article, I introduced you to the world of marketing mix modeling. If you have not read it so far, please do before you proceed. Introduction to Marketing Mix Modeling in Python Which advertising spendings are really driving your sales?towardsdatascience.com There, we have a created a linear regression model that is able to predict sales based on raw advertising spends in several advertising…

Learn about the difference between the frequentist and the Bayesian approach of reasoning — A short Story The three friends Frequentist Frank, Stubborn Stu, and Bayesian Betty go to a funfair where a mysterious-looking tent catches their eyes. Inside, they meet Claire Voyant who claims to be a… fortune teller. The friends don’t believe her, of course — they need proof. So they conduct a little experiment:

Learn how to build MMMs for different countries the right way — Imagine that you work in a marketing department and want to investigate if your advertising investments paid off. People who read my other articles know a way to do this: marketing mix modeling! This method is especially interesting because it also works in a cookieless world. Introduction to Marketing Mix Modeling in Python Which advertising spendings are really driving your sales?towardsdatascience.com

Learn how to solve search problems in a systematic and general way — In the broad field of artificial intelligence, there is the exciting branch of state-space search problems that is filled with all of your favorite puzzles: Sudokus, sliding puzzles, also known as 8-puzzles, 15-puzzles, …, the Tower of Hanoi (recursion lecture flashbacks, anyone?), the pancake sorting problem that we use as…

Create marketing mix models the scikit-learn way — Hello! I noticed that people are really interested in my articles about marketing mix modeling, and that’s why I have created a small present for you: a small library that helps you create simple marketing mix models yourself! In this article, I will show you how to use it. ⚠…

Given the right loss function, a standard neural network can output uncertainty as well — Whenever we build a machine learning model, we usually design it in such a way that it outputs a single number as the prediction. Most models from scikit-learn work like this: tree-based models, linear models, nearest neighbor algorithms, and more. …

Explore a simple reason for why we should deal with causality — One of the later trends in the field of machine learning is the concept of causal inference or causality. In short, it should answer questions like: What is the impact of one feature (called treatment) on another feature (called response or outcome)? Causal inference is by no means a new…

How to evaluate a function and compute an approximation of its gradient in one go — In this article, we will take a look at a recent paper published by Atılım Güneş Baydin, Barak A. Pearlmutter, Don Syme, Frank Wood, and Philip Torr, published on February 17, 2022 — Gradients without Backpropagation. You can find it here. This paper is still under review, so it might…