Context-sensitive application using ML is developed in scope of the lecture Context Sensitive Systems.
This app consists of three parts:
- Record App: App for collecting data from the smartphone accelerometer and upload to edge-ml cloud for three classes:
- Reading: just looking on the smartphone
- Chatting: the same orientation but typing with keyboard
- Calling: Keeping smartphone near the ear and speaking
- Python data analysis with typical ML Pipeline:
- Data Cleaning and Feature Extraction
- Model Selection and Feature Selection
- Optimization of hyperparameters
- Deployment
- sayMyActionApp:
- App for prediction of action using the model from step 2
- To test, go to https://codesandbox.io/s/saymyactionapp-6dx84s?file=/index.html
NB: Tested only on iOS devices