Automatic optimization of outlier detection ensembles using a limited number of outlier examples

Reunanen, Niko; Räty, Tomi; Lintonen, Timo

doi:10.1007/s41060-020-00222-4

Cited by 13 publications

(16 citation statements)

References 76 publications

(182 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Afterward, it re‐weights the training data, allowing the base models to concentrate on patterns misclassified by the previous models. Motivated by Reference 53, the outcomes of the base models are analyzed, then the base models' parameters are tuned using a function to yield optimal performance. The method in Reference 53, however, does not include adaptive base model training.…”

Section: Methodsmentioning

confidence: 99%

“…Motivated by Reference 53, the outcomes of the base models are analyzed, then the base models' parameters are tuned using a function to yield optimal performance. The method in Reference 53, however, does not include adaptive base model training. OAAE's closely related work, ADAHO, 9 includes adaptive base model training but does not take into account anomaly scores margin maximization, which is the focus of OAAE.…”

Section: Methodsmentioning

confidence: 99%

“…The distance‐based algorithm tries to find global anomalies, 57,58 whereas the density‐based algorithm 59 tries to find local anomalies in less dense regions than the distance based algorithm. Motivated by the idea of model optimization 53 and the vital benefits of heterogeneous model formation in SELECT 42 and ADAHO, 9 we combine k‐NN and LOF as base algorithms and adaptively boost their outcomes to demonstrate how different algorithms could be optimized to detect anomalies. Other algorithms, such as the distribution‐based algorithms, were not considered as they assume a distribution 60 .…”

Section: Methodsmentioning

confidence: 99%

“…In order to create a clear contrast between the anomalies and the rest of the data, the anomaly score distributions of the two must have a positive margin 53 . However, the analyzed dataset, here initialized as

{bold-italica}_{bold-italicj}^{boldnorm}

, contains unknown anomalies that prevent accurate margin calculations.…”

Section: Methodsmentioning

confidence: 99%

“…The contrast maximization Equation (7) gives no assurance of diversity of outcomes from the base models. To introduce diversity between the base models' outcomes, we use Equation (7) to account for the correlation and then adjust the optimization to reduce it, similar to Reference 53. Applying Pearson correlation

ρ

shows the dependency between the score vectors

{bold-italica}_{bold-italicj}^{boldnorm} false(bold-scriptD false)

and

a_{bold-italicl}^{bold-italicnorm}

of the

j

th and

l

th base models.…”

Section: Methodsmentioning

confidence: 99%

See 4 more Smart Citations

OAAE: Optimized adaptive anomaly detection ensemble—Base model boosting by parameter optimization

Bii

2021

Engineering Reports

View full text Add to dashboard Cite

Irregular data or anomalies may occur due to human error, miscalculation, or malicious system behavior. The detection of anomalies is a difficult task that requires the use of multiple strategic methods and models. The ideal detection model should assess the strengths and optimize the results of its base models before making final decisions. This task of optimizing the results of the base models contributes to the generation of more accurate results overall. This work presents an optimized adaptive anomaly detection ensemble using heterogeneous algorithms. In the first stage, it adaptively boosts the outcomes of preceding models by weighting their decisions and finding high‐performance ones, and in the second stage, it optimizes the base models by score margin maximization, which enlarges the contrast between the scores of the anomalies and other data prior to fusion to improve detection accuracy. To validate the model, baselines and test results from 10 benchmark datasets are compared. To assess its effectiveness in terms of distinguishing anomalies, the proposed model is tested and evaluated. The analyzed data are presented as cross‐tabulations, with detailed explanations and interpretations. The experiments show an improvement in results even when the least of anomalies (single cases up to 10%) are used.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

{bold-italica}_{bold-italicj}^{boldnorm}

, contains unknown anomalies that prevent accurate margin calculations.…”

Section: Methodsmentioning

confidence: 99%

ρ

shows the dependency between the score vectors

{bold-italica}_{bold-italicj}^{boldnorm} false(bold-scriptD false)

and

a_{bold-italicl}^{bold-italicnorm}

of the

j

th and

l

th base models.…”

Section: Methodsmentioning

confidence: 99%

See 3 more Smart Citations

OAAE: Optimized adaptive anomaly detection ensemble—Base model boosting by parameter optimization

Bii

2021

Engineering Reports

View full text Add to dashboard Cite

show abstract

PSO clustering and pruning-based KNN for outlier detection

2023

View full text Add to dashboard Cite

Fraud Detection in Mobile Payment Systems using an XGBoost-based Framework

2022

View full text Add to dashboard Cite

Mobile payment systems are becoming more popular due to the increase in the number of smartphones, which, in turn, attracts the interest of fraudsters. Extant research has therefore developed various fraud detection methods using supervised machine learning. However, sufficient labeled data are rarely available and their detection performance is negatively affected by the extreme class imbalance in financial fraud data. The purpose of this study is to propose an XGBoost-based fraud detection framework while considering the financial consequences of fraud detection systems. The framework was empirically validated on a large dataset of more than 6 million mobile transactions. To demonstrate the effectiveness of the proposed framework, we conducted a comparative evaluation of existing machine learning methods designed for modeling imbalanced data and outlier detection. The results suggest that in terms of standard classification measures, the proposed semi-supervised ensemble model integrating multiple unsupervised outlier detection algorithms and an XGBoost classifier achieves the best results, while the highest cost savings can be achieved by combining random under-sampling and XGBoost methods. This study has therefore financial implications for organizations to make appropriate decisions regarding the implementation of effective fraud detection systems.

show abstract

Automatic optimization of outlier detection ensembles using a limited number of outlier examples

Cited by 13 publications

References 76 publications

OAAE: Optimized adaptive anomaly detection ensemble—Base model boosting by parameter optimization

OAAE: Optimized adaptive anomaly detection ensemble—Base model boosting by parameter optimization

PSO clustering and pruning-based KNN for outlier detection

Fraud Detection in Mobile Payment Systems using an XGBoost-based Framework

Contact Info

Product

Resources

About