Input Indicators' Optimiztion Needed

I have a fuction that performs calculations based on noisy timestamp data. The function takes 8 inputs, 3 of which are integers and the rest are percentages(floats) and produces a single float result.

I don't have any idea which of those inputs has the greatest effect on the end result of the function. Some of the inputs may have close to zero effect on the end result.

To test the influence of each input I would like to run the function with a basic set of input values with only one indicator taking +/- 1%, -50% and +100% change to its base value. That testing should be done for each of those 8 input indicators.

What's the best Pythonic way to do that?