Phosphorus adsorption model in areas cultivated with sugarcane using data mining techniques

Author(s)

Jeison Andrey Sánchez Parra

• Document type: Master's Dissertation
• Press: Campinas, SP.
• Institution: Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Agrícola
• Defense date: 2020-02-18

Examining board members

Zigomar Menezes de Souza; José Marques Júnior; Edson Eiji Matsura

• Advisor: Stanley Robson de Medeiros Oliveira; Zigomar Menezes de Souza

Abstract

The disposition of macronutrients such as phosphorus in the soil is a determining factor in the growth of plant species. Phosphorus adsorption is characterized by different attributes and natural exposure processes in the soil, which are finally expressed in a database that is difficult to interpret, in which the use of efficient methodologies for knowledge discovery is essential, a data mining is a example. The aim of study was to apply data mining techniques to predict phosphorus adsorption in different relief forms in areas cultivated with sugarcane. The characterization was carried out in agricultural areas of sugarcane in the region of Catanduva, state of São Paulo, in two grids of one hectare with hillslopes cutvature (concave and convex surface). Agricultural practices have been identical over the years, in each area, a sample grid was installed containing 121 points with 10 x 10m spacing, where soil samples were collected to determine physical, chemical, geophysical and mineralogical attributes of the soil. From the database, data mining techniques were used to predict phosphorus adsorption, based on the variables determined in the soil. At the end of the project, an appropriate model was obtained with a high rate of correctness and agreement, with a high predictive potential that allowed decision making in a fast and accurate way with regard to phosphorus adsorption in the soil. In addition, the degree of importance of each analyzed variable was identified, with magnetic susceptibility being the attribute with the greatest influence, and therefore the relationship between the magnetic properties of ferric and aluminum oxide minerals with phosphorus adsorption processes in regions with characteristics similar to soils in southeastern Brazil (AU)

• FAPESP's process: 18/12279-3 - Prediction of phosphorus adsorption in areas cultivated with sugar cane by data mining techniques
• Grantee: Jeison Andrey Sánchez Parra
• Support Opportunities: Scholarships in Brazil - Master