I have two data from two different sensors (Blue = Sensor 1 and Red = Sensor 2) 1.png . Sensor 2 is the gold standard and I wish to recalibrate sensor 1 so as to get an "equalised" data plot like in 2.png (Black). My mentor suggested me using neural Network. My suggestion on where to start reading or incorporating Neural network from ground zero.
Note - I have calibration equation and using these calibration equations for sensor 2 , I get the red plot.
Is the x axis time / sample number?
Why not just do an explicit regression with Sensor 1 as the predictor (IV), and Sensor 2 as the outcome (DV). You may need to try some non-linear models to get a good fit.
Until you have tried these simple solutions, a neural network would seem to be an overly complicate way of implementing a non-linear regression model (which is all it would be)
Dear Arjit Kapoor, if you have errors in both axes, (x and y), I suggest you to use ODR (Orthogonal Distance Regression).
Best regards,
- Badges
- Science topic
- Similar topics
- Psychometrics
- Measurement