A study on classification using machine learning for dementia evaluation

Recently, the number of dementia patients has been increasing due to the aging society. In Japan, a paper-based examination is the main-stream to measure the cognitive function of a subject, but these paper-based tests give much burden to not only patients but also evaluators like facility and medical staff. Therefore, it is necessary to develop a system that can automatically judge the degree of dementia progression, not to burden the doctor. Also, it is required to add play ability not to be a burden on the elderly. From this point of view, the authors developed a recreation game like a puzzle game. This system is easy to play for elderly people and is not a burden. Also, the question-answer is clear, so it is suitable for automatic judgment. We use the obtained features during recreation game to diagnose the degree of dementia progression. We committed the capability of machine learning techniques. Finally, we discussed that the collected features are sufficient to diagnose the degree of dementia progression.