Reinforcement Learning: Life is a Maze


It can be argued that the most important decisions in life are some variant of an exploitation-exploration problem. Shall I stick with my current job or look for a new one? Shall I stay with my partner or seek a new love? Shall I continue reading the book or watch the movie instead? In all of those cases, the question is always whether I should “exploit” the thing I have or whether I should “explore” new things. If you want to learn how to tackle this most basic trade-off read on…
Continue reading “Reinforcement Learning: Life is a Maze”

Teach R to read handwritten Digits with just 4 Lines of Code


What is the best way for me to find out whether you are rich or poor, when the only thing I know is your address? Looking at your neighbourhood! That is the big idea behind the k-nearest neighbours (or KNN) algorithm, where k stands for the number of neighbours to look at. The idea couldn’t be any simpler yet the results are often very impressive indeed – so read on…
Continue reading “Teach R to read handwritten Digits with just 4 Lines of Code”

Understanding AdaBoost – or how to turn Weakness into Strength


Many of you might have heard of the concept “Wisdom of the Crowd”: when many people independently guess some quantity, e.g. the number of marbles in a jar glass, the average of their guesses is often pretty accurate – even though many of the guesses are totally off.

The same principle is at work in so-called ensemble methods, like bagging and boosting. If you want to know more about boosting and how to turn pseudocode of a scientific paper into valid R code read on…
Continue reading “Understanding AdaBoost – or how to turn Weakness into Strength”

Learning R: The Ultimate Introduction (incl. Machine Learning!)


There are a million reasons to learn R (see e.g. Why R for Data Science – and not Python?), but where to start? I present to you the ultimate introduction to bring you up to speed! So read on…
Continue reading “Learning R: The Ultimate Introduction (incl. Machine Learning!)”

Google’s Eigenvector… or how a Random Surfer finds the most relevant Webpages


Like most people, you will have used a search engine lately, like Google. But have you ever thought about how it manages to give you the most fitting results? How does it order the results so that the best are on top? Read on to find out!
Continue reading “Google’s Eigenvector… or how a Random Surfer finds the most relevant Webpages”

Symbolic Regression, Genetic Programming… or if Kepler had R


A few weeks ago we published a post about using the power of the evolutionary method for optimization (see Evolution works!). In this post we will go a step further, so read on…
Continue reading “Symbolic Regression, Genetic Programming… or if Kepler had R”

Learning Data Science: Predicting Income Brackets


As promised in the post Learning Data Science: Modelling Basics we will now go a step further and try to predict income brackets with real world data and different modelling approaches. We will learn a thing or two along the way, e.g. about the so-called Accuracy-Interpretability Trade-Off, so read on…
Continue reading “Learning Data Science: Predicting Income Brackets”

Customers who bought…


One of the classic examples in data science (called data mining at the time) is the beer and diapers example: when a big supermarket chain started analyzing their sales data they encountered not only trivial patterns, like toothbrushes and toothpaste being bought together but also quite strange combinations like beer and diapers. Now, the trivial ones are reassuring that the method works but what about the more extravagant ones? Does it mean that young parents are alcoholics? Or that instead of breastfeeding they give their babies beer? Obviously, they had to get to the bottom of this.
Continue reading “Customers who bought…”