Research on a DBSCAN-IForest Optimisation-Based Anomaly Detection Algorithm for Underwater Terrain Data DOI Open Access

Mingyang Li,

Mingjie Su, Baosen Zhang

et al.

Water, Journal Year: 2025, Volume and Issue: 17(5), P. 626 - 626

Published: Feb. 21, 2025

The accurate acquisition of underwater topographic data is crucial for the representation river morphology and early warning water hazards. Owing to complexity environment, there are inevitably outliers in monitoring data, which objectively reduce accuracy data; therefore, anomalous detection processing key effectively using data. To address anomaly terrain this paper presents an optimised DBSCAN-IForest algorithm model, adopts a distributed computation strategy. First, K-distance graph Kd-tree methods combined determine computational parameters DBSCAN algorithm, applied perform preliminary cluster screening isolated forest subsequently used carry out refined secondary multiple subclusters that were initially screened. Finally, performance verified through example calculations dataset about 8500 points collected from Yellow River Basin, includes both elevation spatial distribution attributes; results show compared with other methods, has greater efficiency outlier detection, rate up 93.75%, parameter settings more scientifically sound reasonable. This research provides promising framework

Language: Английский

Research on a DBSCAN-IForest Optimisation-Based Anomaly Detection Algorithm for Underwater Terrain Data DOI Open Access

Mingyang Li,

Mingjie Su, Baosen Zhang

et al.

Water, Journal Year: 2025, Volume and Issue: 17(5), P. 626 - 626

Published: Feb. 21, 2025

The accurate acquisition of underwater topographic data is crucial for the representation river morphology and early warning water hazards. Owing to complexity environment, there are inevitably outliers in monitoring data, which objectively reduce accuracy data; therefore, anomalous detection processing key effectively using data. To address anomaly terrain this paper presents an optimised DBSCAN-IForest algorithm model, adopts a distributed computation strategy. First, K-distance graph Kd-tree methods combined determine computational parameters DBSCAN algorithm, applied perform preliminary cluster screening isolated forest subsequently used carry out refined secondary multiple subclusters that were initially screened. Finally, performance verified through example calculations dataset about 8500 points collected from Yellow River Basin, includes both elevation spatial distribution attributes; results show compared with other methods, has greater efficiency outlier detection, rate up 93.75%, parameter settings more scientifically sound reasonable. This research provides promising framework

Language: Английский

Citations

1