Algoritmos de aprendizaje automático: Una herramienta para el análisis de datos sobre germinación en semillas

Abstract

One of the challenges facing the agricultural sector for the decade between 2019-2028 is how production will increase and if it will be able to do so in a sustainable way. Historical estimates suggest that agriculture grew proportionally to population (the former tripling and the second doubling since 1960). Despite this, it is estimated that land use for agricultural purposes has increased by only about 10% and is expected to continue. Due to this, various studies have been carried out in recent years around the germination process of seeds, since producing them and being highly viable would help not only to increase agricultural production, but also to the management of plant input. Consequently, the set of statistical tools for its analysis has increased, as well as the complexity of biological data. Even so, the need to contrast these existing tools and techniques is essential, since certain assumptions are usually made in modeling, such as normality in their distribution, independence of these in a sequence, or homoscedasticity of the explanatory variables. Therefore, during the development of this work, machine learning algorithms were used as tools for the analysis of data on seed germination, around the viable development of plants for consumption and aimed at increasing agricultural production.