Research And Application Of Clustering Problem Based On Swarm Intelligence Optimization Algorithms

With the advent of the era of big data,the amount of information generated on the Internet platform every day is the sum of the information in the past decades or even hundreds of years.How to scientifically acquire,store,query,share,analyze and visualize massive data has become an important research topic for researchers.Data mining is a technology to process massive data and extract valuable key information.Clustering analysis is an extremely important part of its technology.As a new heuristic optimization algorithm,swarm intelligence algorithm can well deal with some complex optimization problems.Therefore,the integration of swarm intelligence algorithm and clustering problem has become a front research topic.This paper introduces the related theories of traditional clustering algorithm and swarm intelligence algorithm in detail,analyzes and compares them with traditional algorithms,finds out the algorithm characteristics of different algorithms,such as parameters,usage measures,objective functions,key steps,cycle conditions,summarizes their corresponding defects,and proposes a clustering synthesis algorithm.This paper analyzes and studies the teaching and learning optimization algorithm,and puts forward the teaching and learning optimization algorithm which integrates the niche and non decreasing strategy.Finally,the improved teaching and learning optimization algorithm is used in the density peak clustering algorithm to solve the parameter sensitive problem of the clustering algorithm.Firstly,a clustering algorithm is proposed.The untrained data samples are modeled directly,so that the whole data set can be divided into multiple categories with category labels composed of similar objects.Secondly,we use supervised learning classification to further classify a few existing label data into a classifier,and then use the remaining unlabeled samples to improve the accuracy of the classifier,and then reduce the calculation cost and improve the final data clustering effect.Secondly,a hybrid strategy teaching and learning optimization algorithm is proposed.In order to reduce the possibility of teaching and learning optimization algorithm falling into the optimal solution prematurely and increase the population diversity,a niche selection strategy is introduced after each iteration.Finally,a clustering algorithm is proposed,which integrates teaching and learning optimization algorithm and density difference distance.In order to consider the influence of data point attributes and neighborhood,the Euclidean distance of the original algorithm is replaced by the density difference distance,and the clustering center is selected by the standard difference.Finally,the optimal d_cvalue is found by the algorithm.