Abstract | ||
---|---|---|
This paper analyzes k nearest neighbor classification with training data anonymized using anatomy. Anatomy preserves all data values, but introduces uncertainty in the mapping between identifying and sensitive values. We first study the theoretical effect of the anatomized training data on the k nearest neighbor error rate bounds, nearest neighbor convergence rate, and Bayesian error. We then validate the derived bounds empirically. We show that 1) Learning from anatomized data approaches the limits of learning through the unprotected data (although requiring larger training data), and 2) nearest neighbor using anatomized data outperforms nearest neighbor on generalization-based anonymization. |
Year | Venue | Field |
---|---|---|
2016 | arXiv: Learning | k-nearest neighbors algorithm,Pattern recognition,Best bin first,Computer science,Word error rate,Nearest neighbor graph,Nearest-neighbor chain algorithm,Artificial intelligence,Large margin nearest neighbor,Cover tree,Machine learning,Nearest neighbor search |
DocType | Volume | Citations |
Journal | abs/1610.06048 | 0 |
PageRank | References | Authors |
0.34 | 16 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Koray Mancuhan | 1 | 14 | 3.02 |
Chris Clifton | 2 | 3327 | 544.44 |