A Survey on MR-MNBC: MAX-REL based Feature Selection for the Multi-Relational Bayesian Belief Network
Disha Sheth, Premal Amrishkumar Patel · IJSRD : international journal for scientific research and development · 2015
High dimensional data often contains irrelevant features that reduce the accuracy of data mining techniques and slow down the process and it is hard to interpret so the Feature selection has become an active area in the data mining. Feature selection selects the relevant subset of attributes, provides the better accuracy and improves the comprehensibility of the models. Feature selection also defying the curse of dimensionality to improve prediction performance. We will propose the method which is based on Feature selection as a preprocessing task of MRDM on probabilistic model. We analyzed our algorithm over large pkdd dataset and get the better accuracy compare to the existing methods.