I. Introduction
Lots of future-related information is currently available in text documents. Such information consists of future plans, schedules, predictions, speculations, expectations and so on. The Web or large text collections such as news articles are considered as a reflection of the society. It should be thus possible to harness them for reconstructing collaborative future image held by society. Such kind of social prediction would be generated based on opinions and expectations towards future shared by multiple users according to the “wisdom of the crowds” concept. The results of large scale analysis of future-related information could be useful for individual users in supporting decision making process as well as useful for many domains like futurology, sociological studies, marketing or business intelligence.