Forecasting Crude Oil Price Based on Mutual Information Technique and ANFIS Trained by Biogeography-Based Optimization (BBO) Algorithm  
Author Quang Hung Do


Co-Author(s) Duc Son Nguyen; Thanh Hang Thi Doan; Thu Ha Thi Vu


Abstract Crude oil price forecasting is a crucial task in the field of energy research because crude oil is a world’s major commodity with a high volatility level. This study proposes the Adaptive Neuro-Fuzzy Inference System (ANFIS) with parameters optimized by Biogeography-Based Optimization (BBO) algorithm and mutual information (MI) technique for forecasting crude oil price. The MI is utilized to maximize relevance between inputs and output and minimize the redundancy of the selected inputs. The proposed approach combines the strengths of fuzzy logic, neural network and the heuristic algorithm to detect the trends and patterns in crude oil price data, and thus have been successfully applied to crude oil price forecasting. Other different forecasting methods, including artificial neural network (ANN) model, ANFIS model, and linear regression method are also developed to validate the proposed approach. In order to make the comparisons across different methods, the performance evaluation is based on root mean squared error (RMSE), mean absolute error (MAE), relative absolute error (RAE), root relative squared error (RRSE), and correlation coefficient (R). The performance indexes show that the ANFIS-BBO model achieves lower MAE, RMSE, RAE and RRSE, as well as higher R, indicating that the ANFIS-BBO model is a better method.


Keywords Crude oil price, ANFIS, ANN, BBO, Forecasting
    Article #:  DSBFI19-40
Proceedings of ISSAT International Conference on Data Science in Business, Finance and Industry
July 3-5, 2019 - Da Nang, Vietnam