Application of machine learning and statistics techniques in forecast of biofuel demand

Demand analysis and forecast is fundamental in the strategic planning of the production chain, being significantly important in different segments. The transport sector is a good example, given the high dynamics in biofuel consumption, which requires more intense monitoring of the production, distribution and consumption of this product, in order to prevent failures in supplying population demands. The scenario makes opportunities for the application of predictive statistics and automatic learning techniques, those projections being of great value for understanding the behavior of demand for this resource in the long term. This work exposes different machine learning techniques and statistics, with the purpose of verifying the performance of these techniques in predicting the demand for biofuel. With the aid of the python programming language, the sales data of biofuels, ethanol and biodiesel were used for modeling through three methods: arima, long short-term memory -lstm and gradient boosting. During residue analysis, it was observed that arima models show higher quality in adjustments. However, from the results obtained and through the mape error metric (mean absolute percentage error), the lstm method has the best performance, with a mape error of 11.1% for biodiesel and 11 .3% for ethanol.