Data uncertainty is widespread in a variety of applications. This paper proposes a new Bayesian classification algorithm for classifying uncertain data. In the paper, we apply probability and statistics theory on uncertain data model, and provide solutions for model parameter estimation for both uncertain numerical data and uncertain categorical data. We also prove the correctness of the solutions. The experimental results demonstrate the proposed uncertain Bayesian classifier can be efficiently constructed, and it significantly outperforms the traditional Bayesian classifier in prediction accuracy when data is highly uncertain.
展开▼
