With the advent and dissemination of new media platforms such as audio-based Amazon Alexa, there is growing need to transfer current visual-based information into audio-based information by making it relevant, brief and still easy-to-understand. In this study, we used Textrank algorithm to summarize patient education documents. We first identified sections within each document and apply the algorithm on each section separately. We evaluated our summarization results against manually summarized documents with ROUGE-N metric (40 documents). Average ROUGE F-score for Disease Condition document summarization is 0.454 and average ROUGE F-score for Test Procedure document summarization is 0.413.
Learning Objective: Audience will be able to learn about current gap in the language between physician and patient;
Audience will be able to learn about increasing needs to create brief and relevant information for new media platform;
Audience will be able to learn about Mayo Clinic patient education material content structure;
Audience will be able to learn about TextRank algorithm and its use in document summarization;
Audience will be able to learn about the result of applying TextRank document summarization algorithm.
Yiqing Zhao (Presenter)
Yanshan Wang, Mayo Clinic
Paul Kingsbury, Mayo Clinic
Michael Panzer, Mayo Clinic
Richard Van Ert, Mayo Clinic
Hongfang Liu, Mayo Clinic