Scholarlink

Huge amount of research options over medical text reports do exist in NLP as a domain. The concerned research is based on the text reports of CT scan and MRI tests. Such reports contain lines with some medical abbreviations within them. The fact is, in medical science, one abbreviation can possibly have multiple full forms. So, in case, a line in the report has a medical abbreviation in it, re-generating the line with suitable full form out of multiple possible full forms with “Context Analysis” to increase understanding of the line is the need in medical science. This paper focuses on applying techniques like Cosine Similarity, Jaccard Similarity, usage of pre-trained transformer models like BERT and GPT2 for expansion of the abbreviations by using context analysis. The results of the four techniques are compared with some parameters like Precision, Accuracy, Recall, F1-Score, and BLEU Score.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Contextual Expansion of Medical Abbreviations in Medical Text Reports

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Sakshi Manoj Surve,
  • Hiren Dand

摘要

Huge amount of research options over medical text reports do exist in NLP as a domain. The concerned research is based on the text reports of CT scan and MRI tests. Such reports contain lines with some medical abbreviations within them. The fact is, in medical science, one abbreviation can possibly have multiple full forms. So, in case, a line in the report has a medical abbreviation in it, re-generating the line with suitable full form out of multiple possible full forms with “Context Analysis” to increase understanding of the line is the need in medical science. This paper focuses on applying techniques like Cosine Similarity, Jaccard Similarity, usage of pre-trained transformer models like BERT and GPT2 for expansion of the abbreviations by using context analysis. The results of the four techniques are compared with some parameters like Precision, Accuracy, Recall, F1-Score, and BLEU Score.