|
Verbs can't be classified as positives, negatives etc. They however can be classified into various buckets on the basis on how they behave. What are the various such classifications available? |
|
What you're looking for are called "verb subcategories". Beth Levin's long, detailed compilation of verbs into different classes is pretty classic. See this paper. |
|
One of the approaches is to consider the argument structure -- basically what semantic role a verb assigns to the participants in a sentence. See this paper. |
|
Building on Aditi's answer, much of Beth Levin's work on verb classes has been captured and expanded upon in the VerbNet lexicon. http://verbs.colorado.edu/verb-index/ |
Some verbs can be seen as positive or negative, as least when talking about sentiment analysis. "To disappoint", "to fail", "to excel", etc, are good indicators of sentiment when they occur in text.
I need a generalized classification.
Why exactly do you need such a classification? I don't really understand your question, and it's easier to answer if I understand what is your real problem: how do you plan on using such a classification of verbs?
We are trying to find out what role verbs play in determining the sentiment of a sentence.
Look at this paper http://www.aclweb.org/anthology/N/N10/N10-1120.pdf for an approach to sentiment analysis that incorporates the value and properties of verbs (via a dependency tree). I don't think you're going to find a trivial mapping between preexisting categories of verbs and sentiment. However, you will probably find neutral verbs, positive verbs, negative verbs, generic polarity-"flipping" verbs, and verbs that "flip" the polarity of a sentence only in specific contexts.