I'm using machine learning to do text classification right now, I first use a training data to train my classifier, then use this classifier to classify text document into different classes. With the time goes on, some new words will arise and I need to adjust my training vocabulary to correctly classify these documents, could you please tell me how to do this?
It sounds like you are trying to accomplish a form of online learning: incremental adjustment of the learning parameters using new observations.
These links might be helpful:
Yes it is very important in text mining to use bag of words generated from recent documents.
You will have to create new models every time. Well the time phase depends upon the data generated and your domain. New models can be made on past 4 to 6 months data or in some cases 9 to 12 months data.