USING MACHINE LEARNING TO BUILD AN OPTIMAL MODEL THAT FORECAST THE NET EXPORT OF CEREALS IN ROMANIA

Abstract

The objectives of this article are to build forecasting models that estimate the net exports of several kinds of cereal in Romania. A comparison between the traditional model of regression (the linear model) and machine learning algorithms will be made. Moreover, there will be built an ensemble of these models in order to minimize the RMSE. The implications are complex and represent a mixture of mathematical, econometric and programming skills. The results of the article aimed to identify new methods that increase the confidence level of estimates.