Privacy preservation in data publishing is the major topic of research in the field of data security. Datapublication in privacy preservation provides methodologies for publishing useful information; simultaneouslythe privacy of the sensitive data has to be preserved. There has been little research addressing how toeffectively use the preserved data for data mining in general and for distributed data mining in particular. Here,we propose a new approach for building classifiers using Radial Basis Function (RBF) and Multiple LinearRegression (MLR) by employing sliced data as uncertain data. Use of probability distribution employed in theslicing approach was replaced by classification techniques to enable modeling for sliced data. InRBF, the sliceddata is sent into the input layer, the activation function is executed by the hidden lauer and output layerproduces classified data. In the same manner, MLR calculates approximate value of one or more sliced dataresponses on the basis of certain predictors. Results from the experiments show that these techniques showbetter performance in comparision with other classification approaches.
展开▼
