Selective Colligation and Selective Scrambling for Privacy Preservation in Data Mining

Ishwarya M. V. (Hindusthan Institute of Technology and Science)
K. Ramesh Kumar (Department of Information Technology, Hindusthan Institute of Technology and Science)

Indonesian Journal of Electrical Engineering and Computer Science
Vol 10, No 2: May 2018

The work is to enhance the time efficiency in retrieving the data from enormous bank database. The major drawback in retrieving data from large databases is time delay. This time   hindrance is owed as the already existing method (SVM), Abstract Data Type (ADT) tree pursues some elongated Sequential steps. These techniques takes additional size and with a reduction of speed in training and testing.  Another major negative aspect of these techniques is its Algorithmic complexity. The classification algorithms have five categories. They are ID3, k-nearest neighbour, Decision tree, ANN, and Naïve Bayes algorithm. To triumph over the drawbacks in SVM techniques, we worn a technique called Naïve Bayes Classification (NBC) Algorithm that works in parallel manner rather than sequential manner. For further enhancement we commenced a Naïve Bayes updatable algorithm which is the advanced version of Naïve Bayes classification algorithm. Thus the proposed technique Naïve bayes algorithm ensures that miner can mine more efficiently from the enormous database.