HIGHLIGHTS
SUMMARY
The authors propose a system to automatically detect smoking, alcohol, and drug-use status and the related information, such as amount, frequency, type, quit-time, and period, in medical records. The experimental evaluation on a dataset created and annotated for designing and testing the system, demonstrates the efficiency of the proposed system, achieving an F1-score of up to 0.99 for the detection of substance-use status and up to 0.98 for the detection of substanceuse-related attributes. The authors propose an automated substance-use detection system for clinical text and evaluate . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.