Topic Modeling for International Patients' Consultations Using Natural Language Processing

Sunmi Lee, Sookkyung Sung, Seo Young Kang, Sujin Ha, Eunjoo Jeon

Abstract: We extracted major topic by applying natural language processing and keyword extracting using TF, TF-IDF, TextRank, Yake, KeyBERT. 1452 consultation data were collected from the website and official hospital e-mail. We found six topics categorized into “Medical opinion” related to hospital characteristics and “Non-medical service guidance”. Based on this result, it is necessary to establish marketing plan and develop a digital solution for effective consultation.