Bio-Clinical Named Entity Recognition (NER) Model

This repository contains code for developing and evaluating a bio-clinical Named Entity Recognition (NER) model. The project uses a fine-tuned version of the BERT-based transformer model, pre-trained on a general corpus and adapted for bio-clinical text using a synthetic dataset. The NER model is optimized for recognizing and classifying medical entities in biomedical and clinical datasets.

Project Overview

The project involves:

1. Model Architecture:

   • The model is based on BERT (Bidirectional Encoder Representations from Transformers), a transformer-based deep learning model that has been pre-trained on a large corpus. We fine-tune this model specifically for bio-clinical NER tasks.

2. Synthetic Dataset:

   • The model is trained on a synthetic bio-clinical dataset that mimics real-world clinical scenarios, including entity types such as diseases, symptoms, and treatments.

3. Dataset Preprocessing:

   • Mapping bio-clinical dataset labels into numeric values for model compatibility.
   • Applying data filtration to clean the dataset and ensure high-quality training data.

4. Fine-Tuning:

   • Fine-tuning the pre-trained BERT model on bio-clinical NER tasks allows the model to specialize in recognizing medical entities. This significantly enhances its accuracy on biomedical data.

5. Evaluation:

   • The model is evaluated using a bio-clinical test set, measuring its effectiveness in recognizing clinical entities.

6. Metric Computation:

   • Performance is measured using accuracy, F1 score, precision, and recall to assess the NER model's clinical entity recognition capabilities.