rss
 
 

About

Application Scenario: Future Predictor

Posted February 18, 2010Applications

This scenario implements our future predictor application, built on top of the testbed, which searches for future predictions (plans or estimates) in current documents, tracks them over time and supports reasoning on the impact of such future predictions. 

We will offer the following functionalities in this application scenario:

* Searching the future, i.e., searching current references for explicit mentions of plans or estimates for future events. An example query that can be solved by this application is searching current news reports for estimates of temperature increments due to global warming in 100 years.
* Predicting the future, i.e., inferring new implicit future events based on past occurrences and plans for the future. A usage example for this application is inferring the changes in the price of oil based on current global warming analysis.
* Mining the future, i.e., finding the most important topics associated with a given time segment. This will answer questions such as “what will happen in 2050?”


This application scenario provides functionality for detecting trends even before these trends become apparent in daily events. In addition to offering an important service, this scenario is an excellent framework where most of the research introduced in this proposal will be integrated and evaluated: we will a) detect and retrieve facts and opinions, b) track them over time, c) manage bias and diversity in this vast content space, and d) provide integrated search and navigation capabilities for this multi-dimensional information (i.e., content + bias + time).