Abstract
Evaluation of clustering results (or cluster validation) is an important and necessary step in cluster analysis, but it is often time-consuming and complicated work. We present a visual cluster validation tool, the Cluster Validity Analysis Platform (CVAP), to facilitate cluster validation. The CVAP provides necessary methods (e.g., many validity indices, several clustering algorithms and procedures) and an analysis environment for clustering, evaluation of clustering results, estimation of the number of clusters, and performance comparison among different clustering algorithms. It can help users accomplish their clustering tasks faster and easier and help achieve good clustering quality when there is little prior knowledge about the cluster structure of a data set.
Highlights
Cluster analysis is an important technique in many research areas such as data mining, information science, agriculture technology, and biomedicine
We present a visual cluster validation tool, the Cluster Validity Analysis Platform (CVAP), to facilitate cluster validation
Once clustering results are obtained by a clustering algorithm, the important step is to evaluate clustering solutions to determine an optimal solution or cluster structure for the data set, usually the number of clusters (NC)
Summary
Cluster analysis is an important technique in many research areas such as data mining, information science, agriculture technology, and biomedicine. Once clustering results are obtained by a clustering algorithm, the important step is to evaluate clustering solutions to determine an optimal solution or cluster structure for the data set, usually the number of clusters (NC) This step depends on evaluation of clustering results or cluster validation that aims to find a clustering solution that best fits the given data set. It is usually time-consuming work to accomplish a clustering task because cluster analysis has many aspects to be treated carefully such as data preprocessing, similarity metrics, number of clusters, parameters of clustering algorithms, validity indices, the evaluation of clustering solutions, and so on. We present an efficient cluster validation tool to serve the above purpose
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
