Explain Decision-Tree Training and Clustering Algorithms
Company: Amazon
Role: Data Scientist
Category: Machine Learning
Difficulty: medium
Interview Round: Technical Screen
##### Scenario
Science-breadth interview for an Amazon Applied Scientist role, assessing foundational understanding of common machine-learning algorithms.
##### Question
a) Explain in detail how a decision-tree model is trained: how split points are chosen, how stopping criteria/pruning work, and how overfitting is avoided. b) Name at least three clustering algorithms and describe the core principle behind each one.
##### Hints
Cover impurity measures (Gini/entropy), information gain, pre-pruning vs post-pruning; for clustering mention K-Means, DBSCAN, Hierarchical, Gaussian Mixture, etc., and highlight distance, density or probabilistic assumptions.
Quick Answer: This question evaluates understanding of supervised and unsupervised machine learning techniques, focusing on decision-tree training mechanics (split selection, stopping criteria, pruning and overfitting control) and the core principles behind clustering families such as partitioning, density-based, hierarchical and probabilistic methods.
