Defining custom stop words in NLTK

Hi,
I want to know what is the method for defining our own custom stop word? I'm currently developing a sentiment analysis in my local language in which i'm using Naive Bayes classifier to classify the text. I'm quite new to this type of NLP project so sorry if there's a method that I miss.

Hope you can help me thanks.