Transformer Based Medical Query Matcher

• Tech Stack: Python, scikit-learn, Huggingface, BioBERT, pandas
• Github URL: Project Link

Developed a BioBERT-based Sentence Transformer model fine-tuned for a paraphrase detection task to retrieve the most similar medical queries to a given query. The model was optimized to identify semantic similarities between queries, leveraging BioBERT's domain-specific language understanding.

The project achieved 90% accuracy in retrieving the top 20 most relevant medical queries and 85% accuracy for the top 10, demonstrating the model's effectiveness in medical query retrieval tasks. This solution aids in improving query handling and information retrieval in medical contexts, ensuring accurate and efficient results.