> ML_ALGORITHM // ISOLATION-FOREST_v1.0
Isolation Forest (iForest)
Efficient anomaly detection algorithm that isolates outliers by randomly sub-partitioning feature space using trees.
Ensemble Outlier Detectionclassical-unsupervisedmoderate-posthocmedium (1k-100k)
Back to All AlgorithmsComputational Complexity
Training Complexity:O(trees * subsample * log(subsample))
Inference Complexity:O(trees * log(subsample))
Hardware Profile
CPU Friendly:Yes
Requires GPU:No
Memory Footprint:low
Interpretability & Data
Interpretability Tier:moderate-posthoc
Training Data Needs:medium (1k-100k)
Interpretability Assessment
Anomaly score is a direct monotonic function of average tree path depth.
Suitable Tasks & Supported Modalities
Suitable Tasks:
anomaly detectionoutlier rejection
Supported Modalities:
tabular
Implementing Libraries
Foundational Literature
Isolation ForestFei Tony Liu, Kai Ming Ting (2008) · IEEE International Conference on Data Mining (ICDM)
Common Pitfalls & Warnings
- Axis-aligned splitting creates artifactual low-anomaly regions in empty spaces (addressed by Extended Isolation Forest)