Dementia Detection Using Machine Learning in Males and Females

In this study, we train different Machine Learning classifiers on male and female samples from The ADReSS Challenge dataset. After testing them on data consisting of both male and female samples as well as on data with samples of only the corresponding sex, we use SHapley Additive exPlanations on the models to identify the most important features. Then, we present the results of training the models with only the most important features identified. Finally, we present a comparative analysis based on the observations and insights gathered through these experiments.