Search code examples
mathcomputer-sciencenormalization

How to scale down a range of numbers with a known min and unknown max value


i have a series of experiments, where the outcome of each single experiment is an integer with a known minimum value (which would be the best case) and unknown maximum value. These values are different for every experiment.

Is there a way to normalize such experiments to (as example) [0,1] or any other interval?

The outcome of the experiments is a set of integer c1,c2,...,cn like this:

f(g1)=c1 with 2 <= g1_min <= c1 <= inf
f(g2)=c2 with 2 <= g2_min <= c2 <= inf,

...

f(gn)=cn with 2 <= gn_min <= cn <= inf.

Im looking for a way to visualize the outcome of the experiment and to display how much each cn differs from the respective (optimal) gn_min.

Do you have any idea how to do this?

Greetings :-)


Solution

  • You could play around with something like:

    (1.0 / (x - min + 1.0))
    

    You can take the logarithm of the denominator if you want that kind of scale. The result approaches 1 for infinity. Basic calculus.

    1.0 / (log10(x - min + 1) + 1.0);