MODEL OF DECISION-MAKING EXPERT SYSTEM ON FINANCIAL TIME SERIES

Abstract

Model of expert system, using model indices of two indicators of time series for decision-making is suggested. The first of the indicators defines the trend of series growth, the second one – finds local extremes, using differential derivatives. Parameters of the first indicator are defined by means of digital filtration of input time series. Parameters of digital filter construction are optimized while operation of expert system in real time mode. Efficiency factor of decision-making is suggested, system operation efficiency is evaluated by this factor.