Initial Clustering Based on the Swarm Intelligence Algorithm for Computing a Data Density Parameter

Abstract

To improve the accuracy and efficiency of cluster startup using data density parameters, the author proposes a large data cluster extraction algorithm based on a herd intelligence algorithm. Since clustering to initiate data density parameters is primarily data mining, the author explores data mining clustering based primarily on herd intelligence algorithms. First, the obscure c-key cluster algorithm in the clustering algorithm is analyzed, and then the hybrid jump algorithm in the sub-heuristic herd intelligence optimization technology is optimized in the case of a few parameters by combining the obscure c-means cluster algorithm. The simulation results show that the convergence speed of the fuzzy C-means clustering algorithm and hybrid leapfrog algorithm is slow; the convergence rate of the PSO-FCM algorithm has been improved. Since the fusion algorithm requires fewer adjustment parameters, the cluster centers can be obtained more accurately and quickly with strong robustness and fast convergence. Compared with other algorithms, the fusion algorithm proposed by the author has the best performance in clustering effect, accuracy, convergence rate, and robustness. It is proved that the swarm intelligence algorithm can effectively perform density parameter initialization clustering on computational data.