Second-order difference heuristic model of fuzzy time series based on particle swarm optimization

Fuzzy time series (FTS) has become an effective method for forecasting some typical time series-enrollments, stock price and daily price of foreign exchange-due to its salient capacities for dealing with uncertainty, vagueness in historical observations. However, how to partition the universe of discourse and how to construct fuzzy logic relationships are two unsolved problems which may affect the forecasting accuracy. In order to supply an effective platform, relatively speaking, this paper proposed a novel heuristic model of fuzzy time series based on particle swarm optimization (PSO). The empirical results show that the presented model not only provides more effective fuzzy logic relationship matrix, but obtains higher forecasting accuracy rates than the existing models.