Performance Comparison of Two Classifiers Built by Using Actual and Unrealized Datasets

Decision trees are tree shaped structures that represent sets of decisions. These decisions generate rules for the classification of a dataset.C4.5 is an important classification algorithm. Data security is essential for every data owner. Unrealization approach is based on dataset complementation approach and is an important privacy protecting approach. In this paper, two classifiers, one developed on actual datasets and another on unrealized datasets are evaluated to make performance comparison. Experimental results suggest that classifier built on unrealized datasets gives almost similar results at the same time providing data protection.

Provided by: MIT Publications Topic: Data Management Date Added: Aug 2013 Format: PDF

Find By Topic