Label Noise-Tolerant Hidden Markov Models for Segmentation: Application to ECGs

The performance of traditional classification models can adversely be impacted by the presence of label noise in training observations. The pioneer work of Lawrence and Schölkopf tackled this issue in