Abstract
Oil consumption is one of the main factors that affect industry and economy. Therefore, it is very important to estimate and forecast the consumption of oil. This helps the governments to take the right decisions and avoid the wrong decisions that lead to negative outcomes. For that reason, there are several methods that have been applied to forecast the oil consumption, such as the adaptive neuro-fuzzy inference system (ANFIS) model. It is one of the most popular data mining methods used to perform the forecast. However, the ANFIS model may not be accurate (biased) in all data, since its parameters require to be determined and updated and this may lead to stuck in the local point and not convergence to the optimal value. To this end, this paper presents an alternative oil consumption forecasting method by improving the ANFIS using the sine-cosine algorithm (SCA). In the proposed method called SCA-ANFIS, the parameters of the ANFIS are optimized using the SCA. In order to assess the performance of the proposed SCA-ANFIS method, a real dataset of petroleum products' consumption of three countries, namely, Canada, Germany, and Japan, is used. This dataset is collected on the period between 2007 and 2017, which contains 120 records per month for each country. Moreover, the results of the proposed method are compared with variants of ANFIS models. The experimental results demonstrate that the proposed SCA-ANFIS method outperforms other algorithms.