In case I measured the same substances with two different kits at the same conditions(environment, temperature etc.) and I trained a classification algorithm with one kit's data and than predicted from the other kit's data the result was far from good, in contrast if I trained and predict data from the same kit. What can be the main differences between two development kit? Can it be the manufacturing process, or some internal behaviour of the microcontroller, or anything else? Thanks in advance, Kristóf
... View more