问题
Does there exist a guide with a set of rules for textual analysis / natural language processing?
Do you have some specific developed package (e.g. in Python) for textual sentiment analysis?
Here is the application I am faced with:
Let's say I have two dictionaries, A and B. A contains "negative" words, and B contains "positive" words. What I can do is count the negative and the positive number of words.
This created some issues, such as the following: let's suppose that "exceptionally" is a positive word, and "serious" is a negative word.
If I have the two words following each other, I have "exceptionally serious". In such a case, the two words cancel each other, which means I have 1 negative and 1 positive word. This is not true, because in reality it is a double negative.
So, my question is, is there a set of rules I can apply so that I improve my code, or is there some software that already takes into account such mechanisms, and applies textual sentiment analysis? Is there some implementation which I can feed the dictionaries and provide me with textual sentiment after it applies a set of rules such as double negatives?
回答1:
We did sentiment analysis at San Diego State using nltk with python. Really fun and easy! http://text-processing.com/demo/sentiment/ for an example I entered "exceptionally serious" and it knows that it is NEG.
easy enough example to follow: http://www.laurentluce.com/posts/twitter-sentiment-analysis-using-python-and-nltk/
来源：https://stackoverflow.com/questions/33619635/set-of-rules-for-textual-analysis-natural-language-processing