Abstract:
An approach to forecasting consumer demand based on generalized non-parametric method is developed. Necessary and sufficient conditions for trading statistics to be in a correspondence with the inverse function of demand, satisfying the Law of Demand were found. Warshall-Floyd algorithm verifies these conditions. This algorithm has polynomial complexity over the number of trading statistics points of time. Demand forecasting technique based on the analysis of trading statistics rationalization and the Law of Demand feasibility is suggested.