System of Predicting Dementia Using Transformer Based Ensemble Learning

Abstract

In the previous research of dementia discrimination by voice, a discrimination method using multiple acoustic features by machine learning has been proposed. However, they do not focus on speech analysis in mild dementia patients. Therefore, we proposed a dementia discrimination system based on the analysis of vowel utterance features. The results of the t-test indicated that some cases of dementia appeared in the voice of mild dementia patients. Therefore, we proposed the ensemble discrimination system using a classifier with statistical acoustic features and a Neural Network of transformer models, and the F-score is 0.907, which is the best result.