Phar-LSTM: a pharmacological representation-based LSTM network for drug–drug interaction extraction
Pharmacological drug interactions are among the most common causes of medication errors. Many different methods have been proposed to extract drug–drug interactions from the literature to reduce medication errors over the last few years. However, the performance of these methods can be further impro...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
PeerJ Inc.
2023-12-01
|
| Series: | PeerJ |
| Subjects: | |
| Online Access: | https://peerj.com/articles/16606.pdf |
| _version_ | 1797389083191803904 |
|---|---|
| author | Mingqing Huang Zhenchao Jiang Shun Guo |
| author_facet | Mingqing Huang Zhenchao Jiang Shun Guo |
| author_sort | Mingqing Huang |
| collection | DOAJ |
| description | Pharmacological drug interactions are among the most common causes of medication errors. Many different methods have been proposed to extract drug–drug interactions from the literature to reduce medication errors over the last few years. However, the performance of these methods can be further improved. In this paper, we present a Pharmacological representation-based Long Short-Term Memory (LSTM) network named Phar-LSTM. In this method, a novel embedding strategy is proposed to extract pharmacological representations from the biomedical literature, and the information related to the target drug is considered. Then, an LSTM-based multi-task learning scheme is introduced to extract features from the different but related tasks according to their corresponding pharmacological representations. Finally, the extracted features are fed to the SoftMax classifier of the corresponding task. Experimental results on the DDIExtraction 2011 and DDIExtraction 2013 corpuses show that the performance of Phar-LSTM is competitive compared with other state-of-the-art methods. Our Python implementation and the corresponding data of Phar-LSTM are available by using the DOI 10.5281/zenodo.8249384. |
| first_indexed | 2024-03-08T22:50:42Z |
| format | Article |
| id | doaj.art-78c26514c47448c08fdd8aed088aa727 |
| institution | Directory Open Access Journal |
| issn | 2167-8359 |
| language | English |
| last_indexed | 2024-03-08T22:50:42Z |
| publishDate | 2023-12-01 |
| publisher | PeerJ Inc. |
| record_format | Article |
| series | PeerJ |
| spelling | doaj.art-78c26514c47448c08fdd8aed088aa7272023-12-16T15:05:13ZengPeerJ Inc.PeerJ2167-83592023-12-0111e1660610.7717/peerj.16606Phar-LSTM: a pharmacological representation-based LSTM network for drug–drug interaction extractionMingqing Huang0Zhenchao Jiang1Shun Guo2School of Software Engineering, Shenzhen Institute of Information Technology, Shenzhen, Guangdong, ChinaShenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong, ChinaShenzhen Institute of Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong, ChinaPharmacological drug interactions are among the most common causes of medication errors. Many different methods have been proposed to extract drug–drug interactions from the literature to reduce medication errors over the last few years. However, the performance of these methods can be further improved. In this paper, we present a Pharmacological representation-based Long Short-Term Memory (LSTM) network named Phar-LSTM. In this method, a novel embedding strategy is proposed to extract pharmacological representations from the biomedical literature, and the information related to the target drug is considered. Then, an LSTM-based multi-task learning scheme is introduced to extract features from the different but related tasks according to their corresponding pharmacological representations. Finally, the extracted features are fed to the SoftMax classifier of the corresponding task. Experimental results on the DDIExtraction 2011 and DDIExtraction 2013 corpuses show that the performance of Phar-LSTM is competitive compared with other state-of-the-art methods. Our Python implementation and the corresponding data of Phar-LSTM are available by using the DOI 10.5281/zenodo.8249384.https://peerj.com/articles/16606.pdfPharmacological representationLong short-term memoryMulti-task learningDrug–drug interaction extraction |
| spellingShingle | Mingqing Huang Zhenchao Jiang Shun Guo Phar-LSTM: a pharmacological representation-based LSTM network for drug–drug interaction extraction PeerJ Pharmacological representation Long short-term memory Multi-task learning Drug–drug interaction extraction |
| title | Phar-LSTM: a pharmacological representation-based LSTM network for drug–drug interaction extraction |
| title_full | Phar-LSTM: a pharmacological representation-based LSTM network for drug–drug interaction extraction |
| title_fullStr | Phar-LSTM: a pharmacological representation-based LSTM network for drug–drug interaction extraction |
| title_full_unstemmed | Phar-LSTM: a pharmacological representation-based LSTM network for drug–drug interaction extraction |
| title_short | Phar-LSTM: a pharmacological representation-based LSTM network for drug–drug interaction extraction |
| title_sort | phar lstm a pharmacological representation based lstm network for drug drug interaction extraction |
| topic | Pharmacological representation Long short-term memory Multi-task learning Drug–drug interaction extraction |
| url | https://peerj.com/articles/16606.pdf |
| work_keys_str_mv | AT mingqinghuang pharlstmapharmacologicalrepresentationbasedlstmnetworkfordrugdruginteractionextraction AT zhenchaojiang pharlstmapharmacologicalrepresentationbasedlstmnetworkfordrugdruginteractionextraction AT shunguo pharlstmapharmacologicalrepresentationbasedlstmnetworkfordrugdruginteractionextraction |