Student: Adalberto Andrade
Title: Exploring data analysis to empower forecasting of global fertilizer consumption
Advisors: Pedro Henrique González Silva (advisor) and Eduardo Soares Ogasawara (co-advisor)
Committee: Pedro Henrique González Silva (president), Eduardo Soares Ogasawara (CEFET/RJ), Eduardo Bezerra da Silva (CEFET/RJ), Cristina Gomes de Souza (CEFET/RJ) e Igor Machado Coelho (UFF)
Day/Time: September 23, 2020 / 10h
Fertilizer has received increasing attention from the agribusiness industry, entrepreneurs, governments, and research entities around the world. As critical input for the production chain of food and organic inputs for other sectors, it is important to predict fertilizer consumption, so the
increase in its production could be adequately planned without compromising the environment. It supports decision-making and planning, particularly to agricultural activities, which are strongly dependent on the use of fertilizers. Due that, this research focuses on comparing data analytical approaches to improve predictions of fertilizer consumption under different horizons of steps forward. To do this, We explored ways to optimize the model construction considering different approaches (i.e., pair combinations between data preprocessing and machine learning methods). We evaluate these approaches in a reduced observations set, corresponding to the four main fertilizers of the top ten countries that demand them. The obtained results showed that using the proposed analytic tools can be a promising way to get predictions to plan for future demands.