Improved Particle Swarm Optimization Algorithm And Its Application In Clustering Algorithm

The optimization method is to study how to make a certain(or some)index achieve the optimal discipline under given constraints,and the research of optimization algorithm has always been a key issue in this field.Particle swarm optimization algorithm is a simple and effective algorithm in the optimization algorithm.It combines the individual learning experience and social experience to adjust the evolution direction of the particles,so as to obtain the optimal solution.In the current rapid development of the Internet,the daily increase in the amount of data generated,the data scale jumped from TB to PB or even EB;and data types and data structure is complex,the difficulty of processing increased.At present,the processing and analysis technology of large data is more and more concerned by the government and enterprises,and most of the data mining algorithms are basically the establishment of optimization model,and optimize the objective function(or loss function)with the optimization method to determine the optimal solution.In this paper,an improved particle swarm optimization algorithm is proposed to solve the problem that the particle swarm optimization algorithm is easy to prematurely converge and fall into the local optimal solution,and the improved particle swarm optimization algorithm is applied to the K-means clustering algorithm Data processing platform application.The main work of this paper is as follows:First for easy premature convergence of particle swarm optimization(PSO)and in the shortcoming of local optimum solution,use away from the individual's worst and worst group experience,put forward a kind of far away from the worst of the particle swarm algorithm,and simulated experiment verify the algorithm has good global convergence.Next to the improved particle swarm optimization on Spark cluster parallel programming implementation.Spark platform is currently the most widely used data analysis platform,support Java,Scala,Python,and R and so on the many kinds of language,able to seamlessly combine Hadoop platform,etc.Finally to the improved particle swarm optimization algorithm is applied to the K-means clustering algorithm,for the Iris and Wine data sets,the simulation experiment results is better,and apply it to the telecom positioning blocks,the blocks users get belong to MR information clustering,clustering characteristics after extraction of wireless base stations access between clusters as learning characteristics,in order to later wireless access to the same or similar characteristics of the MR positioning to belong to the building.