Study on Density Clustering based Anomaly Detection for Aquaculture Water
Yu Zhang, Xufeng Hua, Sun Xueliang, Xue Yangyang, Yunchen Tian · 2021
In the process of aquaculture, the quality of water used for aquaculture directly determines the yield and quality of aquatic products. In order to find the abnormal water body in the process of pearl gentian grouper culture, the KANN-DBSCAN algorithm which can adaptively determine the parameters is used to find the anomaly situation, and the results are compared with the culture situation. The results showed that the anomaly cases found by KANN-DBSCAN algorithm were consistent with the culture process, and there were special cases when the data fluctuated. In order to verify the detection effect of the KANN-DBSCAN algorithm, the KANN-DBSCAN algorithm is compared with the LOF algorithm, and the KANN-DBSCAN algorithm is more comprehensive and accurate.