Study And Application Of Adaptive Density Peak Clustering

With the rapid development of modern information science and technology,the importance of data to all walks of life is well known,and there is no doubt that the amount of data in all walks of life is growing rapidly and reaching an unprecedented level.These massive data are from all walks of life,such as business,education,government,scientific research institutions,and networks.Network,or even many off the shelf online or offline databases,and Paul Vientiane,such as text,image,video,audio,animation,hyperlinks and so on.How to mine useful information and understand and make use of information from massive data is particularly critical.Clustering technology plays an important role in data mining.It is used to discover the relationship between data objects in data and estimate the distribution of data.Clustering is to divide a series of data points with similar characteristics into one category or cluster,and the similarity between different cluster classes is as small as possible.Clustering is unsupervised and does not require prior knowledge to guide different categories and how many categories to be classified.Due to wide application,various clustering algorithms have been proposed successively,and different clustering algorithms have different clustering strategies.This paper first introduces the kinds of clustering algorithms,which are suitable for the types of data sets and the advantages and disadvantages of all kinds of clustering algorithms.Then,we focus on the density peak clustering algorithm,and the density peak clustering algorithm is based on the density based clustering algorithm.This algorithm is suitable for processing any shape cluster.A class of data sets.But the density peak algorithm needs to select the cluster center manually,and the decision diagram does not have the function of guiding the classification.We need to test many times on the decision chart and select the best clustering center.So,on the basis of the density peak algorithm,a new clustering algorithm based on density is proposed.We compare the experiments on multiple data sets.We compare the clustering effects of the density peak clustering algorithm and the attractor propagation clustering algorithm on different data and on the different data,including the synthetic standard data set,the text data set,the image data set and the density peak clustering algorithm in these data sets.Excellent.