Problem with encoding

Hi,

I am trying to build a textmining process. I work especially on the Arabic language. by displaying the results I find an encoding problem and characters are illegible.
The corpus is composed of .txt and .html files. and encoding is ANSI.

thank you for your help.