Text mining and Natural Language Processing on Social Media Data giving Insights for Pharmacovigilance: A Case Study with Fentanyl
By: Paulose, R.
Contributor(s): Samy, B. Gopal | Jegatheesan, K.
Publisher: Mumbai Indian Journal of Pharmaceutical Science 2018Edition: Vol. 80 (04) July-August.Description: 762-765.Subject(s): PHARMACEUTICS | Natural language processing | Data mining | Drug abuse | TweetsOnline resources: Click here In: Indian journal of pharmaceutical sciencesSummary: In the present Investigation, the contribution of data mining and natural language processing in pharmacovigilance of fentanyl, a synthetic opioid pain medication was evaluated as a case study. The tweets containing fentanyl as keyword were retrieved from Twitter social media. The tweets were preprocessed in order to make them ready for the analysis. The sentiment analysis algorithm labeled 1927 tweets (41.85 %) as negative, 2067 tweets (44.9 %) as neutral and 610 (13.25 %) tweets as positive. Crisis, dead, death, die, dose, drug, heroin, kill, lethal, opioid, overdose and police were some of the words frequently associated with fentanyl. The high correlation and association of fentanyl with these terms identified by association rule algorithms demonstrated fentanyl abuse and aftermaths in the real world. This study could further be extended to study the region- and population-wise fentanyl misuse and side effects by adding location and user demographic information, which possibly could help in developing drug abuse prevention programs.Item type | Current location | Call number | Status | Date due | Barcode | Item holds |
---|---|---|---|---|---|---|
Articles Abstract Database | School of Pharmacy Archieval Section | Not for loan | 2018308 |
In the present Investigation, the contribution of data mining and natural language processing in pharmacovigilance of fentanyl, a synthetic opioid pain medication was evaluated as a case study. The tweets containing fentanyl as keyword were retrieved from Twitter social media. The tweets were preprocessed in order to make them ready for the analysis. The sentiment analysis algorithm labeled 1927 tweets (41.85 %) as negative, 2067 tweets (44.9 %) as neutral and 610 (13.25 %) tweets as positive. Crisis, dead, death, die, dose, drug, heroin, kill, lethal, opioid, overdose and police were some of the words frequently associated with fentanyl. The high correlation and association of fentanyl with these terms identified by association rule algorithms demonstrated fentanyl abuse and aftermaths in the real world. This study could further be extended to study the region- and population-wise fentanyl misuse and side effects by adding location and user demographic information, which possibly could help in developing drug abuse prevention programs.
There are no comments for this item.