Samuel MonnierA equation-free, geometric proof that the efficient portfolio frontier is a hyperbola.A paper by Merton shows (among other things) that the efficient frontier of modern portfolio theory is a hyperbola, by performing some…Sep 13, 2018Sep 13, 2018
Samuel MonnierAdapted metrics for binary classificationWhen developing a machine learning model, it is important to choose a performance metric adapted to the problem. In this post, I will…Sep 12, 20181Sep 12, 20181
Samuel MonnierCross-validation tools for time seriesDeveloping machine learning models for time series data requires special care, mainly because the usual machine learning assumption that…Sep 7, 20187Sep 7, 20187
Samuel MonnierThe many miracles of string theoryThe content of this blog post should be obvious to anybody reasonably familiar with string theory. Yet I have not seen it written in clear…Aug 26, 20181Aug 26, 20181
Samuel MonnierNotes on the G-research competitionIn this post, I’ll explain my approach to the G-research financial forecasting challenge, a Kaggle-style machine learning competition. I…Apr 18, 2018Apr 18, 2018
Samuel MonnierTruchet and MondrianThis post, originally published here, explains how to construct Mondrian-like patterns algorithmically, and how to use them to create…Apr 15, 2018Apr 15, 2018