k-nearest Neighbor Classification on Spatial Data Streams Using P-trees
PAKDD '02 Proceedings of the 6th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining
An optimized approach for KNN text categorization using P-trees
Proceedings of the 2004 ACM symposium on Applied computing
PARM—An Efficient Algorithm to Mine Association Rules From Spatial Data
IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
Hi-index | 0.00 |
Growth of Internet has led to exponential rise in data communication over the World Wide Web. Several applications and entities such as online banking transactions, stock trading, e-commerce Web sites, etc. are at a constant risk of eavesdropping and hacking. Hence, security of data is of prime concern. Recently, vertical data have gained lot of focus because of their significant performance benefits over horizontal data in various data mining applications. In our current work, we propose a Predicate-Tree based solution for protection of data. Predicate-Trees or pTrees are compressed, data-mining-ready, vertical data structures and have been used in a plethora of data-mining research areas such as spatial association rule mining, text clustering, closed k-nearest neighbor classification, etc. We show how for data mining purposes, the scrambled pTrees would be unrevealing of the raw data to anyone except for the authorized person issuing a data mining request. In addition, we propose several techniques which come along as a benefit of using vertical pTrees. To the best of our knowledge, our approach is novel and provides sufficient speed and protection level for an effective data security.