@nicholdav @vidalthi @BEBischof @ledell That *is* interesting to me! I cursory glance suggests they are doing a modern take on 'forest pruning', which is something i've been thinking about a little recently. I noticed that they cited https://t.co/QLZRUVz
Stumbled on this great idea for "explainable AI" - single decision tree, with a larger hypothesis space by switching from greedy to global optimization. I wonder if real-life perf really does match XGB/LGBM etc. By @dbertsim https://t.co/kERe4UpqaV
A propósito de árboles y machine learning me acordé de esto. No todo es meter a la juguera. https://t.co/sNKU47C4cu
.@dbertsim shows that you can have the interpretability of CART and improved results by finding an optimal tree through mixed-integer programming #orms #ml #PrincetonDay18 #KnightsOfMip Here is the paper: https://t.co/dYuyYLfu1c https://t.co/s4KFdKHKzA
RT @mluebbecke: Optimal classification trees -- #machinelearning aided by #math #optimization > https://t.co/JHOOHgy6Gg
RT @mluebbecke: Optimal classification trees -- #machinelearning aided by #math #optimization > https://t.co/JHOOHgy6Gg
RT @mluebbecke: Optimal classification trees -- #machinelearning aided by #math #optimization > https://t.co/JHOOHgy6Gg
RT @mluebbecke: Optimal classification trees -- #machinelearning aided by #math #optimization > https://t.co/JHOOHgy6Gg
Optimal classification trees -- #machinelearning aided by #math #optimization > https://t.co/JHOOHgy6Gg
Optimal classification trees -- #machinelearning aided by #math #optimization > https://t.co/y69gmMXLJq