I am collecting stats from a number of different sensors on a racing car.
They update every millisecond and are plotted to a real-time graph.
I can see the graph update an observe trends and outliers through a bespoke application.
What I would like to do is extend this bespoke application to automatically alert me to outliers.
To do this I can train the application with 'correct' data based on previous laps completed. If a data-point from a sensor deviates from this correct level on a new lap then I want this to be classified as an outlier (depending on the model I am using).
Any help appreciated.