Yes, it is more like changing the values by good guesses than following some exact procedure.
Following some general advice:
The native output of the neural nets are inherent optical properties (absorption and scattering) of the water. Their magnitude relates to mass concentrations of water constituents, like Chl-a pigment or total suspended matter. The conversion of IOPs into mass concentration depends on the specific properties of the water and may depend on local conditions. How the default values have been derived is mentioned on the C2RCC Community site on the Neural Net page in the section IOP to concentration conversion.
To find the best conversion factor for your area, you need to have in-situ data and compare them with the processing results of C2RCC. You can use the mentioned equation to plot the C2RCC data and vary the factor and exponent values till you found a good match between in-situ data and the result of the equation. Afterwards you can reprocess the data with the new conversion values.
A helpful source on how to do this can be found here: How to Fit an Equation to Data in Excel - EngineerExcel