A feasibility study into alternative pressure air matrass detection through machine learning

Alternating pressure air mattresses play an essential role in the prevention of decubitus. Their presence and smoothless operation
reduces the risk of developing decubitus. Nursing homes in the Netherlands have residents with varying physical conditions, and
have therefore varying types of mattresses. The beds can be equipped with a sensor board that measures pressure and vibrations
continuously. This report presents the investigation in machine learning models that predict the presence of alternative pressure air
mattresses based on the sensor information. Working with a data set of a single nursing home, good accuracy results have been
obtained. Future investigations on explainability of the trained models, as well as testing on larger data sets is needed to improve
the robustness of the model.