Privacy Preserving Naïve Bayesian Classifier Based on Transition Probability Matrix

Recently, lots of researchers pay attention to privacy preserving data mining, which can discover useful knowledge and concurrently preserve the data privacy. In this paper, we propose a Naïve Bayesian classifier based on transition probability matrix, NBCTPM, which uses transition probability matrix to generate the private data. Different from existing techniques, transition probability matrix can also be used for processing non-char data. NBCTPM firstly adopts transition probability matrix to generate private dataset from the original one. Then it builds a Bayesian classifier based on it, and classifies test datasets. Finally, we make some experimental tests on some benchmark datasets. The experimental results show the efficiency of the presented classifier.