HAR-backend is API Deployed in Heroku which gets input values from the accelerometer and gyroscope sensor of the Device and returns the output class label.
The Model used in the API is LSTM .
The Input values/parameter of the API is each (100 x 3) accelerometer sensor , gyroscope sensor and accelerometer due to gravity sensor values.
The API does some pre-processing before predicting the values.
The LSTM model's accuracy is 99% for the given data set.
The possible output class label from the model is ['Biking' ,'Downstairs', 'Jogging', 'Sitting', 'Standing', 'Upstairs', 'Walking'].
API link - API page
