Unsupervised anomaly detection ensembles using item response theory

Ensemble learning combines many algorithms or models to obtain better predictive performance. Ensembles have produced the winning algorithm in competitions such as Netflix Prize. They are used climate modeling and relied upon make daily forecasts. Constructing an ensemble from a heterogeneous set of unsupervised anomaly detection methods presents challenges because class labels ground truth is unknown. Thus, traditional techniques that use cannot be for this task. We Item Response Theory (IRT) – educational psychometrics construct ensemble. IRT’s latent trait computation lends itself can uncover hidden truth. Using novel IRT mapping problem, we downplay noisy, non-discriminatory accentuate sharper methods. demonstrate effectiveness using two real data repositories show it outperforms other techniques. find performs well even if low correlation values.