Conceptual Fuzzy Set Generation Depending on Context

Ambiguity in language is one of the most difficult problems in dealing with word senses using computers. Word senses vary dynamically depending on context. We must specify the context to identify them. We propose here a method to represent such senses using conceptual fuzzy sets. First, we used the modified confabulation model (a prediction method similar to the n-gram model) and word sequences just before the target word to generate atomic conceptual fuzzy sets automatically. Then we generated conceptual fuzzy sets depending on context using the atomic fuzzy sets and a relationship based on cooccurrences. We used a large corpus consisting of 1 million newswire text data in our experiments. The results of these tasks demonstrated that our methodology was effective to generate conceptual fuzzy sets depending on context