A Hybrid Fuzzy Time Series Technique for Forecasting Univariate Data

Alhassan Mohammed B., Muhammad Bashir Mu’azu, Yusuf Abubakar Sha’aban, Shehu Mohammed Yusuf,, Salawudeen Ahmed Tijani & Suleiman Garba


In this paper a hybrid forecasting technique that integrates Cat Swarm optimization Clustering (CSO-C) and Particle Swarm Optimization (PSO) with Fuzzy Time Series (FTS) forecasting is presented. In the three stages of FTS, CSO-C found application at the fuzzification module where its efficient capability in terms of data classification was utilized to neutrally divide the universe of discourse into unequal parts. Then, disambiguated fuzzy relationships were obtained using Fuzzy Set Group (FSG). In the final stage, PSO was adopted for optimization; by tuning weights assigned to fuzzy sets in a rule. This rule is a fuzzy logical relationship induced from FSG. The forecasting results showed that the proposed method outperformed other existing methods; using RMSE and MAPE as performance metrics.             

Full Text:



  • There are currently no refbacks.