Concetti Chiave
Toronto Metropolitan University's approach to TREC Clinical Trials Track 2023.
Sintesi
Directory:
Abstract and Introduction
Utilization of advanced natural language processing techniques and neural language models for clinical trials retrieval.
Shift from traditional recruitment methods to questionnaire-based identification of suitable clinical trials.
Methodology
Data preparation involving XML parsing and feature extraction.
Information retrieval and document ranking using Sentence Transformer and Doc2Vec models.
Results and Evaluation
Submission of four runs with varying models, highlighting the performance of Sentence Transformer model.
Conclusion and References
Key Highlights:
Transition to questionnaire-based clinical trial identification.
Use of PubMed Parser for data extraction.
Application of neural language models for trial retrieval.
Comparison between Sentence Transformer and Doc2Vec models.
Performance evaluation based on NDCG scores.
Statistiche
試験応募の結果、4つのランが提出されました。
表1は、提出されたランのNDCGスコアを示しています。
表2は、提出されたランの評価結果を示しています。