How modify WEKA source code to support persian text?(utf8)

Question

I want to use result of weka classification and cluster in my program. How can I use persian(utf8) data in weka? I use TextDirectoryLoader to load my text files and when I want to store my data in .csv format it doesn't save in correct format.

Can anybody help me?

Answer 1

you should change RunWeka.ini file in weka folders .you can use this solution to fix your problem: http://weka.8497.n7.nabble.com/Reading-utf8-arff-files-td15810.html

Answer 2

var loader=new weka.core.converts.TextDirectoryLoader();
loader.setcharset("UTF-8");