Medical Term Normalization Data for Social Media from EMNLP 2015

Data and supplementary information for the paper 'Adapting Phrase-based Machine Translation to Normalise Medical Terms in Social Media Messages' published at EMNLP 2015. The dataset likely contains a collection of phrases from tweets related to adverse drug reactions, used to map laymen's terms to medical concepts. The research was conducted by Nut Limsopatham at the University of Cambridge.

Use Cases

• Training phrase-based machine translation models for medical term normalization based on the described approach.
• Evaluating word vector similarity methods for mapping social media phrases to medical concepts.
• Benchmarking text normalization techniques for adverse drug reaction monitoring in social media.

Strengths

• Dataset is associated with a peer-reviewed paper published at EMNLP 2015, a major NLP conference.
• The described method combining phrase-based MT and word vectors reportedly outperformed baselines by up to 55%.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

University of Cambridge, author Nut Limsopatham.

Collection Method

Likely collected from social media platforms such as DailyStrength and Twitter, focusing on phrases related to adverse drug reactions.

Time Range

null

Freshness

Data is associated with a 2015 conference paper; freshness unverified.

Geography

null