ADHD-ML

Head and Coordination

Team

Funding

Duration

11/2013 - 06/2014

Project ADHD-ML

Machine Learning for Detecting Hyperactivity in Children

We investigated intelligent methods to predict hyperactivity in children using accelerometer data. We considered the classification of recorded time-series as well as segmentation tasks.

Johnson et al. (2012) studied hyperactivity in children using magnetic resonance imaging together with machine learned classifiers. Although the reported classification rates are very promising, their approach is not only financially costly but also expensive in terms of time and it does not allow measurements in the everyday life of the children. In contrast to resource intensive MRI experiments, accelerometers serve as conventional and cheap sensors and have become the method of choice for measuring physical activity.

In this project we studied the detection of hyperactivity in children using accelerometer data. Machine learning techniques, such as support vector machines, were used for learning and inference. The goal was to provide an inexpensive and simple alternative to detect hyperactivity.

Selected Publications

Brefeld, U., & Scheffer, T. (2006). Semi-supervised Learning for Structured Output Variables. Proceedings of the International Conference on Machine Learning.

Fernandes, E. R., & Brefeld, U. (2011). Learning from Partially Annotated Sequences. Proceedings of the European Conference on Machine Learning.

Gawrilow, C., Kühnhausen, J., Schmid, J., & Stadler, G. (2014). Hyperactivity and motoric activity in ADHD: Characterization, assessment, and intervention. Frontiers in Psychiatry, 5. doi:10.3389/fpsyt.2014.00171