Skip to main content

Table 10 Error analysis of the false negatives by the MyC classifier trained on the hepatobiliary disease data set with 70% under-sampling

From: Improving sensitivity of machine learning methods for automated case identification from free-text electronic medical records

Type of error

N (%)

Evidence not in the model

13 (38)

Evidence removed by negation/speculation filter

12 (35)

Spelling variations

5 (15)

Labeling error

4 (12)