Study, Density-based Clustering Algorithm

Along with the rapid development of modern information technology such as the Internet, people must face massive information every day. How to extract useful information from massive data has increasingly become a hot topic of concern. In the domain of data mining, cluster analysis is an important research problem. As a basic means of information processing, cluster analysis technology has become people's concern in recent years. Cluster analysis has also gained a wide range of research and application in machine learning, pattern recognition, data mining, information retrieval and many other fields.The clustering algorithm mainly includes partition-based clustering algorithms, hierarchical clustering algorithm, density-based clustering algorithm, grid-based clustering algorithm and model-based clustering algorithm. Among that the most primary virtue of the clustering algorithm relies on a density-based is to discover clusters of arbitrary sharp as well as to distinguish noise. However, these algorithms require input parameters. Sensitivity to parameters input is one disadvantage. Other disadvantage is that finding cluster in the data with non-uniform density is very difficult. But actually, it is depend on itself that object belong to certain cluster. So, we could find a method to find the values of parameters.Attribute of object reflect distances to another object, especially to the latest distance. So we could count these distances, and analysis frequency of these distances. From the frequency histogram of latest distance, we could find the useful information. Image segmentation method could divide up the gray histogram of image. To frequency histogram of latest distance, it is the same. From these thresholds, it is easy to cluster to all objects. In this paper, an algorithm based on threshold segmentation and density is proposed.The primary reseaarch include as follows:1. Show some density-based clustering algorithms and its meliorative algorithms, and analyse the problem of primary density-based clustering algorithms.2. Introduce image segmentation method simply, and analyse the technique of image segmentation in image process in detail. Especially, analyse the threshold segmentation of based gray histogram.3. Discuss calculation of the frequency histogram of latest distance, and assimilate the techniaue of image segmentation in image process to divide the histogram of lastest distance. So these thresholds could apply to density-based clustering algorithms, and treat with the data respectively.