Improved Naive Bayes with Mislabeled Data
This addresses labeling mistakes in real-world text classification applications, but it is incremental as it builds on the Naive Bayes method.
The paper tackles the problem of mislabeled data in text classification by proposing an improved Naive Bayes method, which uses an EM algorithm to optimize a log-likelihood function based on a specified incorrect label generation mechanism, and results show it greatly improves performance over standard Naive Bayes with mislabeled data.
Labeling mistakes are frequently encountered in real-world applications. If not treated well, the labeling mistakes can deteriorate the classification performances of a model seriously. To address this issue, we propose an improved Naive Bayes method for text classification. It is analytically simple and free of subjective judgements on the correct and incorrect labels. By specifying the generating mechanism of incorrect labels, we optimize the corresponding log-likelihood function iteratively by using an EM algorithm. Our simulation and experiment results show that the improved Naive Bayes method greatly improves the performances of the Naive Bayes method with mislabeled data.