Improved Adaptive Density Clustering Algorithm

The purpose of data mining from large amounts of data is to find out the potential and useful information. In face of large amounts of data, the priority is to cluster it properly. The set of physical or abstract data objects can be divided into similar classes which contain all the data objects, and this process is called clustering. Clustering techniques mainly include:partitioning methods, hierarchical method, the methods based on density and methods based on grid etc. With the development of computer power and the improving complexity of business, the types of the data will be more and more, and the data will be more and more complex. As a result, data mining will play a growing role.Clustering analysis based on density is of great significance in data mining. However, Most of the clustering algorithms in clustering process need to input parameters, and the inputting parameters have great effects on the cluster results. Obviously, the parameters chosen by the users of algorithms may not suitable to the database. By all of this, we say the parameters we input are sensible to the algorithms. Aimed at this shortage, on the basis of the clustering algorithms based on density and cluster algorithms based on hierarchy, we give a algorithm named ADCA in which the parameters needed are produced on the basis of the data objects themselves. Through the analysis of the experimental results, we are sure that we are able to get reasonable clustering results.