Random Forests, Made Simple!!!

Random Forests, Made Simple!!! What is a random forest, and how does it beat a single decision tree? One decision tree is jumpy and dramatic - but a hundred of them, voting together, become the most reliable predictor in all of machine learning. In this video: the wisdom of crowds, bootstrapping, random splits, majority voting, out-of-bag error, variable importance, and why more trees can never hurt. Chapters: 0:00 Hello and welcome 0:17 The route for this lesson 0:34 Fixing the jumpy tree 0:54 Norm's question: a hundred times the overfitting? 1:15 Average many noisy guesses 1:37 Bootstrapping: 100 remixes 1:58 Force the trees to be different 2:19 The crowd votes 2:45 Let's vote for real: 3-2, PASS 3:06 Out-of-bag: a free test set 3:28 What matters: studied #1 3:52 More trees can't hurt 4:11 The price: accuracy for interpretability 4:32 Forests do numbers too 4:51 Random forests in the wild 5:11 Two dials, both forgiving 5:35 Bagging's big picture - next: k-nearest neighbors 5:59 In summary 6:24 Wrap-up and the next lesson 🔔 Subscribe to Practical Data Science for statistics and machine learning, made simple. Keep practicing! #statistics #machinelearning #datascience #randomforests