Intuitionistic fuzzy inference system with weighted comprehensive evaluation considering standard deviation-cosine entropy: a fused forecasting model

Recent development in intuitionistic fuzzy inference system (IFIS) has been emerged with promising results in defining uncertain information and improving its capacity to forecast real-world time series data. Nonetheless, many factors such as non-linearity data, stochastic dynamic problems and weights of attributes are explicitly affect the performance of IFIS. In this paper, we introduce a new method of determining weight of variable to perform an intuitionistic fuzzy comprehensive evaluation that to be fused with an IFIS. In order to weight the credibility of each causal variable in the experimental of particulate matter (PM10) data, a synthesized weight that is established from two different methods of weighting is developed. Two objective weightings known as the intuitionistic fuzzy-standard deviation and intuitionistic fuzzy-cosine entropy are combined as to consider statistical properties and trigonometric properties within the intuitionistic fuzzy set environment. This paper also investigates whether the two weighting methods have the same impact on the forecasting. The experimental results show that our proposed synthesized weighting method outperforms other three weight methods in PM10 forecasting under IFIS environment. The experimental results also verify that different methods of weighting have different influence on performance of the forecasting. This is the first identifiable synthesized weighted comprehensive evaluation that fused in IFIS and its application to PM10 forecasting. Finally, some consideration regarding the limitations of the study and potential research direction is presented.