A feature selection method for classification of ADHD

At present, the classification of brain diseases through neuroimaging data is a hot topic. Attention deficit hyperactivity disorder (ADHD) is usually diagnosed by the standard scale. However, the traditional diagnostic methods have high misdiagnosis rate and time consuming. In this paper, we discussed the classification of ADHD by using the feature subset obtained by preprocessing and feature selection of fractional amplitude of low-frequency fluctuation (fALFF) in resting-state functional magnetic resonance imaging (rs-fMRI) data. We proposed a feature selection algorithm based on Relief algorithm and verification accuracy (VA-Relief). The experimental results show that fALFF can be used to realize the high accuracy classification of ADHD by using our feature selection algorithm and preprocessing method. Therefore, it is possible to use rs-fMRI data and machine learning methods to assist the diagnosis of brain diseases.