Comparar métodos

	Detecção Fora da Distribuição	Isolation Forest	Calibração de Modelo
Área	Aprendizado de máquina	Aprendizado de máquina	Aprendizado de máquina
Família	Machine learning	Machine learning	Machine learning
Ano de origem	2017	2008	2017
Autor original	Hendrycks & Gimpel	Liu, F.T., Ting, K.M. & Zhou, Z.-H.	Platt; Guo et al.
Tipo	Reliability and safety method for neural networks	Unsupervised ensemble (random partitioning trees)	Post-hoc probability correction technique
Fonte seminal	Hendrycks, D., & Gimpel, K. (2017). A baseline for detecting misclassified and out-of-distribution examples in neural networks. International Conference on Learning Representations.	Liu, F.T., Ting, K.M. & Zhou, Z.-H. (2008). Isolation Forest. IEEE ICDM, 413–422.	Guo, C., Pleiss, G., Sun, Y., & Weinberger, K. Q. (2017). On calibration of modern neural networks. International Conference on Machine Learning, 1321–1330.
Outros nomes	OOD Detection, Novelty Detection, Open-Set Recognition, Dağılım Dışı Tespit	Isolation Forest (Aykırı Değer Tespiti), iForest, isolation forest anomaly detection	Classifier Calibration, Probability Calibration, Score Calibration, Model Kalibrasyonu
Relacionados	3	5	3
Resumo	Out-of-Distribution (OOD) detection is a set of techniques that identify when a deployed machine learning model receives inputs that differ significantly from its training data distribution. Introduced as a formal problem by Hendrycks and Gimpel in 2017, these methods enable models to flag unfamiliar inputs rather than silently produce unreliable predictions, making them foundational to trustworthy and safe AI deployment in high-stakes domains.	Isolation Forest is an unsupervised machine-learning method for anomaly and outlier detection, introduced by Liu, Ting and Zhou in 2008, that isolates anomalies through random partitioning of the data. It works without any labelled anomaly data and scales to high-dimensional datasets.	Model calibration is a post-hoc technique that adjusts the probability outputs of a trained classifier so that predicted confidence scores match empirical outcome frequencies. A classifier is said to be perfectly calibrated if, among all predictions made with confidence p, exactly a fraction p of them are correct. Systematic miscalibration of modern deep neural networks was rigorously documented by Guo et al. (2017), who showed that networks trained with standard cross-entropy loss tend to be overconfident, and proposed temperature scaling as a simple, effective remedy.