A Machine Learning Model for Accurate Cardiopulmonary Resuscitation Chest Compression Feature Extraction

The National Centre for Biotechnology Information claims that 83% of in-out patient died due to complication and lack of proper CPR performance. Currently there are only subjective or qualitative CPR parameter measurements which lead to inaccuracy. Existing hardware solutions may lead to neurological deficits and other complications. The proposed solution develops an accurate machine learning model using k-means algorithm to extract the features of CPR bio signals for different health groups from infants to senior adults. When the model is passed on to Simulink interfaced with dolorimeter, it predicts the amount of loading (the pressure needed to be applied on the chest) and unloading (the depressure) that the clinician or the CPR metrician should follow to save the live.