DTU: A Decision Tree for Uncertain Data
PAKDD '09 Proceedings of the 13th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining
| Hi-index | 0.00 |
Mining for clusters in a database with uncertain data is a hot topic in many application areas, such as sensor database, location database, face recognition system and so on. Since it is commonly assumed that most of the objects which are contained in a high-dimensional dataset are located in a low-dimensional subspace, mining clusters in a subspace in an uncertain database is a new task. In this paper, we adopt and combine fractal correlation dimension with fuzzy distance function to find out the clusters in a low-dimensional subspace in an uncertain database. We also propose the fuzzy kth NN algorithm to retrieve the kth nearest neighbor which can accelerate the process of mining. The experiments show that the new algorithm works well in an uncertain database.