Word prediction using WordNet

Lets say I have a text in English and there is a word missing somewhere in it. I have a list of candidate words from a dictionary with no other information. These candidate words are selected by some other, rather inaccurate, algorithm. I would like to use WordNet and the context around the missing word to assign probabilities to candidate words.

There is an obvious ad-hoc way that came to my mind on how to solve this. One way would be to extract "interesting" words surrounding the missing word, calculate semantic similarity with every candidate word according to some metric and assign probabilities to candidate words based on the average score.

However I was unable to find any useful research papers on this problem.

So, what i'm asking is if you're aware of any research (papers) about this problem, how do you find my proposal and do you have a better idea?

---

You can start from Experiments: Enriching indirect answers. A good article is Semantic web access prediction using WordNed.

链接地址: http://www.djcxy.com/p/62408.html

上一篇: FrameNet WordNet映射

下一篇: 使用WordNet进行Word预测