Affinity propagation clustering algorithm based on large-scale data-set

Affinity Propagation (AP) algorithm is not effective in processing large-scale data-sets, so the paper purposed an affinity propagation clustering algorithm based on large scale data-set, called LD-AP. First, we use the idea of grid clustering to divide large data-sets into small datasets and running AP in them to ensure the center of clustering. Then, we introduced the structure similarity matrix to calculate the distance of the cluster center. At last, we used Density peak Clustering Algorithm (DP) algorithm to cluster the cluster again. The experimental results show that the improved algorithm is better than the original algorithm in the clustering effect and computation speed.