Event representation in fuzzy neural networks with qualitative inputs

There is no satisfactory method for allowing a new event, such as troop movements, aircraft spottings or the news of an earthquake, to influence the findings of a fuzzy neural network with qualitative inputs. Yet in some applications, such as pattern recognition, situation assessment, or financial trend prediction, it is essential to combine the event data with input data. The input data may be certainty of events that are likely to occur and/or qualitative description of the events that have already occurred. The author provides a solution to this problem by aggregating the certainty of the new event with the "object to a class belongingness" of a fuzzy neural network having qualitative inputs.