Which SageMaker algorithm detects irregularities in unlabeled patient data?
A healthcare company wants to detect irregularities in patient vital signs that could indicate early signs of a medical condition. The company has an unlabeled dataset that includes patient health records, medication history, and lifestyle changes. Which algorithm and hyperparameter should the company use to meet this requirement?
Community Insight
The question tests matching data characteristics (unlabeled, irregularity detection) to an unsupervised anomaly-detection algorithm — the trap is reaching for XGBoost because the data is tabular.
Unlabeled health records with no anomaly labels call for an unsupervised anomaly-detection algorithm, and Amazon SageMaker AI Random Cut Forest fits that need with num_trees controlling the ensemble size. This page explains why RCF, not XGBoost, k-means, or DeepAR, is the correct choice (D).
Picking XGBoost with a large max_depth because the dataset is tabular; XGBoost is supervised and needs labels, and max_depth above 100 would simply overfit rather than reveal irregularities.
Community Discussion (2 comments)
Comments & Corrections
No comments yet — spotted an error or have a note? Share it below.
Expert Analysis
Why the Answer Is Correct
The dataset is explicitly unlabeled and the goal is to surface "irregularities" in patient vital signs, which is the textbook definition of unsupervised anomaly detection. Amazon SageMaker AI Random Cut Forest is purpose-built for this: it builds an ensemble of random cut trees and outputs an anomaly score per record, so no ground-truth labels are needed during training. Among its hyperparameters, num_trees sets how many trees make up the forest, and raising it above 100 stabilizes the anomaly scores and reduces variance. That combination of algorithm and tuning directly satisfies both halves of the requirement.
Why the Other Options Are Wrong
XGBoost (A) is a supervised gradient-boosted tree algorithm that needs a labeled target column; with no labels there is nothing to train against, and setting max_depth above 100 is a recipe for severe overfitting, not a way to "regulate tree complexity." k-means (B) is a clustering algorithm that partitions records into k groups of similar points — it assumes roughly spherical clusters, requires you to guess k in advance, and produces cluster assignments rather than per-record anomaly scores. DeepAR (C) is a supervised time-series forecasting algorithm that needs historical target values per series and outputs probabilistic forecasts, not anomaly flags, so epochs tuning is irrelevant here.
Community Comment Notes
Both commenters converge on RCF for the same reason: as one learner put it, Random Cut Forest is chosen because the question asks to detect "irregularities," which maps to anomaly detection. Another commenter (Jojopoopoo) points out that XGBoost suits supervised prediction on tabular data, so the unlabeled nature of this dataset makes Random Cut Forest "more suitable." Their reasoning matches the exam's intent: classify the learning problem first (unsupervised), then pick the algorithm family that matches it.
Official Reference
Exam Strategy
Before reading the options, tag the scenario on two axes: labeled vs. unlabeled, and prediction vs. anomaly vs. forecast. Any answer using a supervised algorithm (XGBoost, DeepAR) on unlabeled anomaly data can be eliminated immediately, and absurd hyperparameter values like max_depth > 100 are usually planted distractors.
Frequently Asked Questions
Why is k-means clustering wrong for detecting irregular vital signs?
k-means only partitions records into k similar groups and needs k chosen upfront; it produces cluster membership, not the per-record anomaly score required to flag irregular patients.
What does the num_trees hyperparameter control in SageMaker Random Cut Forest?
It sets how many random cut trees form the forest. Larger values (well above 100) reduce variance in the anomaly scores at the cost of more compute and memory.
Related Analysis
Practice All MLA-C01 Questions
Access 115 questions with complete answers and detailed explanations.
View Full MLA-C01 Practice Test →