History
In which decade did Tin Kam Ho publish a paper describing random decision forests?
90s (1995).
Card 1 of 259. Answer: 90s (1995).
90s (1995).
Card 1 of 259. Answer: 90s (1995).
About · Quizzes · History of AI · To suggest new questions or report an error send me a dm.
The 1990s. Tin Kam Ho's paper "Random Decision Forests" appeared in 1995 in the proceedings of the Third International Conference on Document Analysis and Recognition (ICDAR) in Montreal, written at AT&T Bell Labs. A random forest is an ensemble of decision trees whose votes or averaged outputs form the prediction.
Ho's idea was to grow each tree on a randomly chosen subset of the features, the random subspace method, which she expanded in a 1998 IEEE TPAMI paper. Leo Breiman's 2001 paper "Random Forests" combined random feature selection at each split with bagging, giving every tree a bootstrap sample of the rows. The randomness decorrelates the trees: averaging B trees with variance σ² and pairwise correlation ρ gives variance ρσ² + (1 − ρ)σ²/B.
Interviewers follow up with why adding trees does not overfit, what max_features controls, and how out-of-bag rows give a free validation estimate. Do not credit it all to Breiman: Ho introduced random decision forests, and Breiman built the modern algorithm and popularized the name.