Pergunta de entrevista da empresa C3 AI

MCQ: 1. whether linear/logistic regression error has to be normal. 2. In a RF with n features, n_sub for each tree, only 1 feature significant. How many partitions do not consider the feature. 3. Roll a die until sum>=61. What's the most likely sum. 4. prior (uniform) & post bad rate when sampling 30, 3 are bad. Coding: longest substring where each character at least appears k times