Privacy-Protected KNN Classification Algorithm Based on Negative Database
Negative database (NDB) is a new type of information representation method that protects private data by storing data in the raw data complement set. The KNN classification algorithm is a classic classification algorithm, and the Euclidean distance formula is one of the most commonly used distance calculation formulas in classification algorithms. However, the distance calculation method for the existing KNN classification algorithm based on negative database is the one-hot coded Hamming distance formula. For this encoding method, when data set have many attributes, the length of the binary string becomes extremely long after encoding, thereby it increases the computational cost and complexity of the classification algorithm. In this paper, we proposed a KNN classification algorithm based on the Euclidean distance formula on the negative database, which is used to complete the classification research under the premise of protecting data security. The experimental results show that the algorithm in this paper achieves high classification accuracy.
KeywordsNegative database KNN classification Euclidean distance
- 1.Zhan, J.Z., Chang, L., Matwin, S.: Privacy preserving k-nearest neighbor classification. IJ Netw. Secur. 1(1), 46–51 (2005)Google Scholar
- 5.Esponda, F., Trias, E.D., Ackley, E.S., Forrest, S.: A relational algebra for negative databases. University of New Mexico, Technical Report (2007)Google Scholar
- 6.Esponda, F., Forrest, S., Helman, P.: Enhancing privacy through negative representations of data. NEW MEXICO UNIV ALBUQUERQUE DEPT OF COMPUTER SCIENCE (2004)Google Scholar
- 7.Zhao, D., Luo, W., Liu, R., Yue, L.: A fine-grained algorithm for generating hard-toreverse negative databases. In: 2015 International Workshop on Artificial Immune Systems (AIS), pp. 1–8. IEEE (2015)Google Scholar
- 11.Dheeru, D., Karra Taniskidou, E.: UCI machine learning repository (2017). http://archive.ics.uci.edu/ml. Accessed 10 June 2018