Naive Bayes Classifier for Noisy Medical Information Dataset

• Activity Number: 521
• Type: Contributed
• Date/Time: Thursday, August 11, 2005 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Xiaowei Yang*+ and Yirong Yang
• Companies: BayesSoft, Inc. and BayesSoft, Inc.
• Keywords: naive Bayes classifier ; noisy data ; classification ; Bayesian network

Abstract

Classification is one of the major tasks in knowledge discovery and data mining. Naive Bayes classifier, in spite of its simplicity, has proven surprisingly effective in many practical applications. In real datasets, noise is inevitable because of the imprecision of measurement or privacy-preserving concerns. In this paper, we develop a new approach for learning the underlying naive Bayes classifier from noisy observations. Our method, based on linear equation systems and statistical analysis mechanisms, reconstructs the underlying probability distributions of the noise-free dataset from the observed noisy data. By incorporating the noise model into the learning process, we improve the classification accuracy. Furthermore, as an estimate of the underlying naïve Bayes classifier for the noise-free dataset, the reconstructed model can be combined easily with new observations corrupted at different noise levels to obtain a good predictive accuracy. We apply our approach on both synthetic and real application dataset, especially on medical information dataset.