Using artificial intelligence to classify breathing data of Long Covid patientsby Computational Neuroscience

Topic: Using artificial intelligence to classify breathing data of Long Covid patients

As part of a DFG-project, we recorded breathing data from a healthy population and a population of Long Covid patients in a rebreathing task. The Long Covid patients were divided into two subgroups depending on their symptoms: one group indicated perceived breathlessness, the other group did not. You will train machine learning algorithms to recognize the distinguishing features of the control group and both Long Covid subgroups and then automatically classify new breathing data into these subgroups. Using cross-validation, you will calculate the classification accuracy of the resulting algorithm. For this project, you will need to analyse the pre-processed breathing data and corresponding labels indicating meta data from the participants, perform literature research on potential classification approaches and implement the most promising approaches. With your help, we hope to learn more about distinguishing features in the data and gain new insights on differences in breathing patterns between controls and the two patient groups.

Main topics:
-    Feature extraction and classification with machine learning 
-    Breathing data analysis
-    Programming and data analysis in Matlab
-    Literature research on classification approaches of human physiological data

For further questions, please contact Prof. Dr.-Ing. Stefan Glasauer at stefan.glasauer@b-tu.de