A Modified K-Medoid Method to Cluster Uncertain Data Based on Probability Distribution Similarity

Provided by: International Journal Of Engineering And Computer Science
Topic: Data Management
Format: PDF
Clustering on uncertain data, one of the essential tasks in data mining. The traditional algorithms like K-Means clustering, UK-Means clustering, and density based clustering etc, to cluster uncertain data are limited to using geometric distance based similarity measures, and cannot capture the difference between uncertain data with their distributions. Such methods cannot handle uncertain objects that are geometrically indistinguishable, such as products with the same mean but very different variances in customer ratings. In the case of K-medoid clustering of uncertain data on the basis of their KL divergence similarity, they cluster the data based on their probability distribution similarity.

Find By Topic