Heartivity is about using our research to monitor the electrocardiogram (ECG) and the seismocardiogram (SCG) with chips and accelerometers. Infer with Machine Learning the activity that is being performed with the data collected.

​

We take the raw data obtained from our in house study and perform feature extraction on the x, y and z-axis accelerometer data and filtering and smoothing techniques on the ECG & SCG signals. Then we plugged in the processed data into our Machine Learning algorithm.

​

Our Machine Learning technique uses 3 different classifiers Decision Tree, K-Nearest Neighbor and Naïve Bayes. The algorithm then performs training and testing via the 80/20 split train test and 25 cross-fold validation on our data.