The option of creating classes is especially helpful for combined datasets. Classes allow to group data based on a text wave. Often this is used e.g. for data that combines measurements from different size fractions, stations or instruments. Classes do not only allow to plot only part of the data for analysis but also to perform PMF on certain classes or to give different relative errors when performing PMF.
Classes are created under “Select and/or modify effective PMF input”. Under “classes” new classes can be defined. They can either be defined via the drop-down menu from a text wave that is stored under “root:SoFi:External_data:Raw_input” or via indices that can be directly entered. If the classes are loaded from a wave, the wave must be a text wave with the same dimesion as the PMF input time series or profile, respectively. For more detail check the manual chapter 5.4.1.1.
