A two-factors and multi-orders self-adaptive fuzzy time series model based on fuzzy logical relationships trees and particle swarm optimization

Fuzzy time series have been widely used to deal with forecasting problem. In this paper we propose a two-factors and multi-orders self-adaptive fuzzy time series forecasting model based on fuzzy logical relationships trees and particle swarm optimization (PSO) algorithm. First, we propose a algorithm with weight to partition the universe of discourse. Then we group the fuzzy logical relationships trees for each factor and define the corresponding heuristic rules with forecasting weights for forecasting. We apply PSO algorithm to track the optimal value of above weights. Specifically, the robustness of lower orders fuzzy logical relationships and the precision of higher orders fuzzy logical relationships are exploited simultaneously to improve the forecasting accuracy. The trading data of Taiwan capitalization weighted stock index (TAIEX) and Dow Jones Indexes are used as benchmark data for training and testing, and the experimental results show that the proposed model gets better forecasting performance.