NLM launched the Consumer Health Information and Question Answering project to provide consumers with reliable health information. A consumer spelling tool (CSpell) was developed for spelling error correction in the NLP pipeline due to the high spelling error rate in consumer questions. A systematic approach was developed to retrieve consumer health terminology (CHT) from the UMLS Metathesaurus/MEDLINE. After adding CHT to the default dictionary in CSpell, we observed a 6% improvement in F1.
Learning Objective 1: Learning the importance of consumer spelling correction tool in the consumer NLP systems and how to utilize consumer health terminology to improve the system performance.
Learning Objective 2: Learning the importance of consumer health informatics, consumer health terminology, and consumer NLP systems.
Chris Lu, NIH/NLM/LHNCBC
Dina Demner-Fushman (Presenter)