I am about to use Weka for processing a dataset that contains numeric and nominal values. As Persian and Arabic are the secondary languages I am using in my Windows 7 operating system, I assumed that this might be the reason that when I am trying to save a loaded CSV data file in Weka as an ARFF file format, the numbers are all saved as question (?) marks! However, even by removing these languages from the Control Panel setting, nothing changed.
Moreover, I have upgraded my Java version from 8 to 9 recently. I am not sure if this could be the reason for this.
I searched for a probable reason on the internet, though could not find any solution. Thanks, everybody in advance.