Unbalanced big data density clustering method based on dynamic grid
Aiming at the problem of cumbersome clustering and low accuracy of clustering results in imbalanced big data,a density clustering method based on dynamic grids is proposed.By dividing the dynamic grid and setting a threshold for the corresponding grid density,adaptive grid generation is carried out to achieve the corresponding density clustering effect.The algorithm monitors the abnormal trajectories of users through sample training and testing,proposes the concept of class similarity to partition different lattice clusters,and detects noise as abnormal data to ensure the comprehensiveness of data detection.After actual experimental verification,the improved algorithm has better processing effect and higher accuracy for problems such as imbalanced big data.
dynamic gridunbalanced big datadata streamsimilar classesanomaly trajectories
NETL
NSTL
万方数据
