Handedness tests for preschool children: A novel approach based on graphics tablets and support vector machines

摘要

Handedness is a certain kind of laterality, namely the preference of humans for a certain hand (the dominant hand). Reliable handedness tests are needed in various contexts, for example, to avoid a wrong writing education of children. In this article, we propose a new approach for a gradual rating of hand proficiency which consists of three subtests and is suited for preschool children. We demonstrate the benefits of using a graphics tablet for handedness tests, investigate the advantages of the three subtests and their combination, and outline the possibility of interpreting the gradual output of a classifier. The classification is based on ensembles of support vector machines that are trained with sample data. Input of these classifiers are attributes that reflect various aspects of hand motor skills. The most relevant input attributes of the classifiers are selected from a large set of possible attributes with a ranking technique based on the Gini index. We evaluate this approach using a data set with data gathered from 53 preschool children aged between five and six and a half (45 with certain and known handedness, 8 with uncertain or ambiguous handedness).