fuzzy logic, deduction

Time series prediction is an important task in many areas including economics, biology etc. There are many standard methods to solve this task like Winter's method for time series prediction. These methods have one essential imperfection - they are not transparent. It means if a prediction is produced it cannot be observed why and how such an result has been obtained or such explication of results is described in hardly-readable mathematical relations. We present a new method which overcomes this imperfection and it is based on already very successful formalism of fuzzy linguistic IF-THEN rules presented and developing by the group of Prof. Novak.