An NLP-Inspired Data Augmentation Method for Adverse Event Prediction Using an Imbalanced Healthcare Dataset
This paper proposes a data augmentation method for imbalanced healthcare datasets. This method was inspired by a data augmentation method in natural language processing (NLP) that generates synthetic sentences for training by replacing some words with similar words. The proposed method generates syn...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9845410/ |
| _version_ | 1811309119408177152 |
|---|---|
| author | Tomoki Ishikawa Takahiro Yakoh Hisashi Urushihara |
| author_facet | Tomoki Ishikawa Takahiro Yakoh Hisashi Urushihara |
| author_sort | Tomoki Ishikawa |
| collection | DOAJ |
| description | This paper proposes a data augmentation method for imbalanced healthcare datasets. This method was inspired by a data augmentation method in natural language processing (NLP) that generates synthetic sentences for training by replacing some words with similar words. The proposed method generates synthetic patient records by replacing patient backgrounds with similar backgrounds. In this paper, the cosine similarity of the distributed representations was used as the similarity metric between patient backgrounds. The distributed representations of the patient backgrounds were generated by the skip-gram model. To confirm the performance improvement with the proposed data augmentation method, the prediction performance of adverse events (AEs) caused by drug administration was experimentally evaluated on a real-world medical dataset with 1,510,137 records. The combination of the proposed data augmentation method and a conventional undersampling method resulted in an 80.0% improvement in accuracy and a 40.0% improvement in the precision and F1-score. The multifaceted evaluation demonstrated that the proposed method is effective, especially for predicting AEs with positive ratios ranging from 1.0% to 2.1%, which are difficult to predict with conventional machine learning methods but should be predictable in the medical field. |
| first_indexed | 2024-04-13T09:36:34Z |
| format | Article |
| id | doaj.art-650992bc01ef471892d400c171af0c6f |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-04-13T09:36:34Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-650992bc01ef471892d400c171af0c6f2022-12-22T02:52:05ZengIEEEIEEE Access2169-35362022-01-0110811668117610.1109/ACCESS.2022.31952129845410An NLP-Inspired Data Augmentation Method for Adverse Event Prediction Using an Imbalanced Healthcare DatasetTomoki Ishikawa0https://orcid.org/0000-0002-5811-3994Takahiro Yakoh1https://orcid.org/0000-0002-7031-3286Hisashi Urushihara2https://orcid.org/0000-0001-6913-9930Graduate School of Science and Technology, Keio University, Yokohama, JapanDepartment of System Design Engineering, Keio University, Yokohama, JapanDivision of Drug Development and Regulatory Science, Keio University, Tokyo, JapanThis paper proposes a data augmentation method for imbalanced healthcare datasets. This method was inspired by a data augmentation method in natural language processing (NLP) that generates synthetic sentences for training by replacing some words with similar words. The proposed method generates synthetic patient records by replacing patient backgrounds with similar backgrounds. In this paper, the cosine similarity of the distributed representations was used as the similarity metric between patient backgrounds. The distributed representations of the patient backgrounds were generated by the skip-gram model. To confirm the performance improvement with the proposed data augmentation method, the prediction performance of adverse events (AEs) caused by drug administration was experimentally evaluated on a real-world medical dataset with 1,510,137 records. The combination of the proposed data augmentation method and a conventional undersampling method resulted in an 80.0% improvement in accuracy and a 40.0% improvement in the precision and F1-score. The multifaceted evaluation demonstrated that the proposed method is effective, especially for predicting AEs with positive ratios ranging from 1.0% to 2.1%, which are difficult to predict with conventional machine learning methods but should be predictable in the medical field.https://ieeexplore.ieee.org/document/9845410/Adverse event predictiondata augmentationdistributed representationhealthcare datasetimbalanced dataset |
| spellingShingle | Tomoki Ishikawa Takahiro Yakoh Hisashi Urushihara An NLP-Inspired Data Augmentation Method for Adverse Event Prediction Using an Imbalanced Healthcare Dataset IEEE Access Adverse event prediction data augmentation distributed representation healthcare dataset imbalanced dataset |
| title | An NLP-Inspired Data Augmentation Method for Adverse Event Prediction Using an Imbalanced Healthcare Dataset |
| title_full | An NLP-Inspired Data Augmentation Method for Adverse Event Prediction Using an Imbalanced Healthcare Dataset |
| title_fullStr | An NLP-Inspired Data Augmentation Method for Adverse Event Prediction Using an Imbalanced Healthcare Dataset |
| title_full_unstemmed | An NLP-Inspired Data Augmentation Method for Adverse Event Prediction Using an Imbalanced Healthcare Dataset |
| title_short | An NLP-Inspired Data Augmentation Method for Adverse Event Prediction Using an Imbalanced Healthcare Dataset |
| title_sort | nlp inspired data augmentation method for adverse event prediction using an imbalanced healthcare dataset |
| topic | Adverse event prediction data augmentation distributed representation healthcare dataset imbalanced dataset |
| url | https://ieeexplore.ieee.org/document/9845410/ |
| work_keys_str_mv | AT tomokiishikawa annlpinspireddataaugmentationmethodforadverseeventpredictionusinganimbalancedhealthcaredataset AT takahiroyakoh annlpinspireddataaugmentationmethodforadverseeventpredictionusinganimbalancedhealthcaredataset AT hisashiurushihara annlpinspireddataaugmentationmethodforadverseeventpredictionusinganimbalancedhealthcaredataset AT tomokiishikawa nlpinspireddataaugmentationmethodforadverseeventpredictionusinganimbalancedhealthcaredataset AT takahiroyakoh nlpinspireddataaugmentationmethodforadverseeventpredictionusinganimbalancedhealthcaredataset AT hisashiurushihara nlpinspireddataaugmentationmethodforadverseeventpredictionusinganimbalancedhealthcaredataset |