Data mining retrieves useful information from various sources like internal and external data. At the same time providing security to the data to be mined and also to the patterns extracted from the dataset is a challenging task, which could be accomplished by formulating various privacy models. This work provides rotation based transformation method for preserving privacy of data during data publishing and it is tested by various most familiar classification techniques such as Artificial Bee Colony(ABC), Decision Tree(DT), Bayesian network, Artificial neural network(ANN) and Probabilistic neural network(PNN). The suitable classification technique has been recommended based on the classification accuracy. The developed methodology attempts to perturb the original data values that fit within the security range. The proposed method is evaluated by comparing the data mining results of the perturbed data values with the original data values. © Research India Publications.