Abstract: Aiming at the problem of abnormal data mining in marine Argo buoy monitoring data,an anomaly detection algorithm based on distance criterion was proposed based on the improved K-means algorithm.The algorithm redefines the proximity of ocean data,blocks according to the size and distribution of the data,and adaptively selects alternative initial clustering centers.In the iterative process of the algorithm,using the distance mean of the data objects in the cluster relative to the clustering center,the global consideration is given to the data objects in the cluster according with the abnormal features to detect the anomalies.The simulation dataset and the real dataset are verified by experiments,and the comparison results show that it is superior to the contrast algorithm in clustering performance and anomaly detection.

Key words: Anomaly detection, Argo buoy data, Block, K-means algorithm, Proximity