Using NLP to assign (Buy/Sell/Indeterminate) label to a paragraph of text related to a stock?

I'm working in python and I was thinking as a rudimentary approach to just have a dictionary of words which would correspond to each label and whichever has the most matches in the paragraph would be the label assigned (perhaps with weights assigned to each word as well based off of frequency in a training set?).

Are there any better approaches that aren't extremely difficult to implement? I know there's the NLTK library for python and I've gone through the tutorial but nothing really jumped out at me on how to use it to do something like this.

I was also thinking maybe some kind of ML classifier using centroids or something along those lines, but I wouldn't even know where to begin with quantifying the text..

edit: I should mention that I'm trying to determine what the text is advocating, not trying to assign a label based off information about the stock. In other words, taking an article about a stock which (in most cases) has a position on whether to buy or sell, and determining which position that is.