| Summary: | 碩士 === 慈濟大學 === 醫學資訊學系碩士班 === 104 === The ability to use comics to present health education information has been proved to enhance recall frequency of contents for viewers. It can also increase viewers’ interest and motivation in health education information. However, drawing comics requires talent and may take time for health care practitioners. How to quickly generate health education information in comics is a worth thinking issue. To solve this problem, this study proposes a method that adopts text sentiment analysis and develops an information system, in order to help health care practitioners quickly convert text into comics strips. Our study presents a method combining term frequency–inverse document frequency (TF-IDF) method based on a text dictionary of emotional expressions. We also propose an algorithm that can automatically update the data in the dictionary, by using the statistical method to find commonly used emotion words. In the experiment, we evaluate this system by calculating the accuracy of emotion detection with ten-fold cross-validation approach. The system provides a new way for medical personnel to generate the information in healthy education in less time. Consequently, this system can speed up the production of health education content and increase its quantity.
|
|---|
