Built-in Algorithm Breakdown Matrix

Matrix 1 of 2

Algorithm Name	Machine Learning Type	Core Purpose / What it Measures	When to Use (Exam Trigger Keywords)	Real-World Predictive Maintenance Example
Random Cut Forest (RCF)	Unsupervised Anomaly Detection	Assigns a numeric anomaly score to each data point based on how much it deviates from the normal baseline.	Predicting equipment failures Sensor data streams Spike / outlier detection	Monitoring a wind turbine's vibration sensor to flag sudden, unusual shaking before a blade breaks.
Linear Learner	Supervised Classification / Regression	Fits a straight line or decision boundary to predict a continuous number or a binary category (1 or 0).	Binary classification Predicting labeled outcomes Classify status as healthy/at-risk	Reviewing a dataset where engineers manually labeled rows as 1 (Broken) or 0 (Working) to predict future states.
DeepAR	Supervised Time-Series Forecasting	Uses recurrent neural networks to predict future values across thousands of related timelines over a specific window.	Demand forecasting Predicting future continuous numbers Supply chain / inventory optimization	Estimating exactly how many spare machine parts your warehouse will need to order over the next 6 months.
K-Means	Unsupervised Clustering	Groups data points into a specific number (k) of distinct clusters based on geometric distance.	Audience/Customer segmentation Group similar patterns Discover hidden structures	Sorting a fleet of 1,000 trucks into 3 distinct operational profile groups (e.g., Long-Haul, City-Delivery, Idle-Heavy).