Samuele MazzantiinTowards Data ScienceCausality in ML Models: Introducing Monotonic ConstraintsMonotonic constraints are key to making machine learning models actionable, yet they are still quite unusedSep 612Sep 612
Samuele MazzantiinTowards Data ScienceWhat’s Wrong With R-Squared (And How to Fix It)Even if you think you are using R-Squared out-of-sample, you are not. Here is whyAug 718Aug 718
Samuele MazzantiinTowards Data ScienceForget Statistical Tests: A/B Testing Is All About SimulationsHow simulations outperform traditional stats in that they are easier to understand, more flexible, and economically meaningfulJul 414Jul 414
Samuele MazzantiinTowards Data ScienceHypothesis Testing Explained (How I Wish It Was Explained to Me)Most resources focus on things like Confidence and Power. But they don’t really matter: here is what you should care aboutMay 136May 136
Samuele MazzantiinTowards Data ScienceWhy You Should Never Use Cross-ValidationIn real-world applications, using randomized cross-validation is always a bad choice. Here is why.Mar 2740Mar 2740
Samuele MazzantiinTowards Data ScienceAre Outliers Harder To Predict?An empirical analysis about whether ML models make more mistakes when making predictions on outliersFeb 411Feb 411
Samuele MazzantiinTowards Data Science“Approximate-Predictions” Make Feature Selection Radically FasterFeature selection is so slow because it requires the creation of many models. Find out how to make it blazingly faster thanks to…Nov 17, 202322Nov 17, 202322
Samuele MazzantiinTowards Data ScienceYour Dataset Has Missing Values? Do Nothing!Models can handle missing values out-of-the-box more effectively than imputation methods. An empirical proofOct 9, 20237Oct 9, 20237
Samuele MazzantiinTowards Data ScienceWhich Features Are Harmful For Your Classification Model?How to calculate the Error Contribution of the features of a classifier, with the goal of understanding and improving the modelSep 12, 202311Sep 12, 202311
Samuele MazzantiinTowards Data ScienceYour Features Are Important? It Doesn’t Mean They Are Good“Feature Importance” is not enough. You also need to look at “Error Contribution” if you want to know which features are beneficial for…Aug 21, 202315Aug 21, 202315