System Project for Ukrainian-language Feedback Tonality Analysis in the Health Care Field Based on BERT Model

Authors:

Abstract
The medical field is always popular, especially now in Ukraine in conditions the war. The developed technology for analyzing the tonality of Ukrainian-language reviews in the field of health care based on the BERT model should potentially help to eliminate a number of problems that exist in medical institutions in Ukraine: long expectation, medical specialists appointments, poor treatment of patients. The subsystem that will use natural language processing tools is integrated into the system project as the auxiliary one. Its function should inform management about emergency cases recorded in reviews. There are many analogues, but mostly for processing foreign languages, this subsystem should primarily be aimed at the Ukrainian language. The BERT model for analyzing the tonality of Ukrainian-language reviews has proved to be accurate in 10 out of 11 examples (more than 90%), while the only example when the model hasn’t predict the correct answer is when the review is sufficiently complex and unique, when trying to analyze others, even more complex reviews, the model performed well.

This publication has 14 references indexed in Scilit: