Outlier Detection and Explanation Method Based on FOLOF Algorithm DOI Creative Commons
Lei Bai, Jia-Sheng Wang, Yu Zhou

и другие.

Entropy, Год журнала: 2025, Номер 27(6), С. 582 - 582

Опубликована: Май 30, 2025

Outlier mining constitutes an essential aspect of modern data analytics, focusing on the identification and interpretation anomalous observations. Conventional density-based local outlier detection methodologies frequently exhibit limitations due to their inherent lack preprocessing capabilities, consequently demonstrating degraded performance when applied novel or heterogeneous datasets. Moreover, computation factor for each sample in these algorithms results considerably higher computational cost, especially case large This paper introduces a method named FOLOF (FCM Objective Function-based LOF) through examination existing algorithms. The approach starts by applying elbow rule determine optimal number clusters dataset. Subsequently, FCM objective function is employed prune dataset extract candidate set outliers. Finally, weighted algorithm computes degree anomaly set. For analysis, Golden Section was used classify underlying causes outliers can be revealed exploring properties point factors dimension property. has been validated artificial datasets, UCI dataset, NBA player demonstrate its effectiveness.

Язык: Английский

Outlier Detection and Explanation Method Based on FOLOF Algorithm DOI Creative Commons
Lei Bai, Jia-Sheng Wang, Yu Zhou

и другие.

Entropy, Год журнала: 2025, Номер 27(6), С. 582 - 582

Опубликована: Май 30, 2025

Outlier mining constitutes an essential aspect of modern data analytics, focusing on the identification and interpretation anomalous observations. Conventional density-based local outlier detection methodologies frequently exhibit limitations due to their inherent lack preprocessing capabilities, consequently demonstrating degraded performance when applied novel or heterogeneous datasets. Moreover, computation factor for each sample in these algorithms results considerably higher computational cost, especially case large This paper introduces a method named FOLOF (FCM Objective Function-based LOF) through examination existing algorithms. The approach starts by applying elbow rule determine optimal number clusters dataset. Subsequently, FCM objective function is employed prune dataset extract candidate set outliers. Finally, weighted algorithm computes degree anomaly set. For analysis, Golden Section was used classify underlying causes outliers can be revealed exploring properties point factors dimension property. has been validated artificial datasets, UCI dataset, NBA player demonstrate its effectiveness.

Язык: Английский

Процитировано

0