Cost-sensitive Naïve Bayes Classification of Uncertain Data

Data uncertainty is widespread in real-word applications. It has captured a lot of attention, but little job has been paid to the research of cost sensitive algorithm on uncertain data. The paper proposes a novel cost-sensitive Naïve Bayes algorithm CS-DTU for classifying and predicting uncertain datasets. In the paper, we apply probability and statistics theory on uncertain data model, define the utility of uncertain attribute to total cost, and propose a new test strategy for attribute selection algorithm. Experimental results on UCI Datasets demonstrate the proposed algorithm can effectively reduce total cost, and significantly outperforms the competitor.