SentenceSegmenter-MIMIC Open-source Model - Free for Sentence Segmentation and Annotation of MIMIC-III Clinical Records

Sentencesegmenter MIMIC

Developed by dongfangxu

This model is based on BiomedBERT and is used for sentence segmentation of MIMIC-III clinical records, predicting BIO annotations.

Sequence Labeling

Safetensors

EnglishOpen Source License:MIT #Clinical Text Segmentation #BIO Annotation #MIMIC-III Specialized

Downloads 14.96k

Release Time : 4/22/2025

Model Overview

This model is a token classification model specifically designed for processing clinical text, predicting sentence boundaries (B for sentence beginning, I for sentence interior, O for sentence exterior).

Model Features

Clinical Text Optimization

Optimized for biomedical and clinical text based on BiomedBERT.

BIO Annotation

Accurately identifies sentence boundaries using the BIO annotation scheme.

MIMIC-III Training

Trained on the MIMIC-III clinical records dataset, suitable for real-world clinical data.

Model Capabilities

Clinical Text Processing

Sentence Boundary Detection

Sequence Labeling

Use Cases

Clinical Record Processing

Electronic Health Record Analysis

Automatically segments sentences in clinical records for subsequent information extraction and analysis.

Relevant performance metrics were reported in the EMNLP 2024 paper.

🚀 Clinical Sentence Segmentation Model

This model is designed for sentence segmentation of MIMIC - III clinical notes, offering accurate BIO tagging for efficient text processing.

🚀 Quick Start

This model is tailored for the sentence segmentation of MIMIC - III notes. It processes clinical text as input and predicts BIO tagging. Here, 'B' marks the beginning of a sentence, 'I' represents the inside of a sentence, and 'O' denotes the outside of a sentence.

✨ Features

High - precision Segmentation: Capable of accurately segmenting clinical sentences from MIMIC - III notes.
BIO Tagging: Provides clear BIO tagging for each token in the text.

📚 Documentation

Model Description

This model is used for sentence segmentation of MIMIC - III notes. It takes the clinical text as input and predict BIO tagging, where B indicates the Beginning of a sentence, I represents Inside of a sentence, and O denotes Outside of a sentence. More details of this model is in the paper [Automatic sentence segmentation of clinical record narratives in real - world data](https://aclanthology.org/2024.emnlp - main.1156/). The sample code of using this model is at github.

Our segmentation model is based on [microsoft/BiomedNLP - BiomedBERT - base - uncased - abstract - fulltext](https://huggingface.co/microsoft/BiomedNLP - BiomedBERT - base - uncased - abstract - fulltext), and we trained on MIMIC - III notes for a sequence labeling (token classification) task.

Property	Details
Model Type	token classification model
Language(s) (NLP)	en
Parent Model	[microsoft/BiomedNLP - BiomedBERT - base - uncased - abstract - fulltext](https://huggingface.co/microsoft/BiomedNLP - BiomedBERT - base - uncased - abstract - fulltext)
Resources for more information	GitHub Repo

📄 License

This project is licensed under the MIT license.

📄 Citation

Dongfang Xu, Davy Weissenbacher, Karen O’Connor, Siddharth Rawal, and Graciela Gonzalez Hernandez. 2024. [Automatic sentence segmentation of clinical record narratives in real - world data](https://aclanthology.org/2024.emnlp - main.1156/). In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 20780–20793, Miami, Florida, USA. Association for Computational Linguistics.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご