Banca de DEFESA: DEYVID DIEGO CARVALHO MARANHÃO

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : DEYVID DIEGO CARVALHO MARANHÃO
DATE: 20/09/2023
TIME: 14:00
LOCAL: INSTITUTO DE ZOOTECNIA
TITLE:

Machine learning applied to beef livestock in Brazil.

KEY WORDS:

Random forest, RFE, livestock, beef cattle.

PAGES: 68
BIG AREA: Ciências Agrárias
AREA: Zootecnia
SUMMARY:

There will be a substantial increase in the world's population in the coming years and consequently a growing demand for food. Agriculture is responsible for most of what is consumed, being a key activity in the context of food security. Within the ecosystem that involves these activities are agriculture and livestock. Although both are somewhat correlated, within the context in question, livestock derived from the production of beef cattle requires special attention, as it is one of the main sources of protein for human beings. The production of beef cattle in Brazil is carried out in the most diverse ways and conditions, with a diversity of social, environmental, economic, management, food supply, genetics, among other issues involved, which reflects in complex interactions, which can be decisive in the success of the activity. These interactions represent a challenge when dealing with planning, sometimes limiting us in terms of the lack of insights that can be decisive in decision making. Therefore, artificial intelligence can be an important tool in order to understand these interactions. Thus, the objective of this study was to identify the main variables associated with beef cattle production in Brazil using the Randon Forest algorithm. Finally, it is expected that data manipulation will contribute to the planning of the bovine production chain, becoming a disruptive innovation in the context of Brazil, providing information that will help strategies of rural organizations and public policies, which are important in the agribusiness ecosystem. It was verified that the areas where limestone application practices and/or other soil correctives are not used, represent the most important variables in relation to the number of cattle produced. These aspects suggest that the largest portions of the areas fall short of what is expected in terms of management practices. Practices related to pasture management are not considered a priority, which justifies the high rates of degradation related to livestock activity, which consequently interferes with the biomass produced and animal support. Cattle farming is considered a villain with regard to soil degradation and deforestation, aspects corroborated in this study. Therefore, a commitment on the part of the government is necessary, with regard to public policies aimed at rural extension and scientific research showing the benefits related to practices, especially justifying through economic metrics the returns obtained since it employs management techniques from soil. According to the evaluations, the production of flour represents greater importance in relation to the amount of beef produced in the agroindustry. Next, variables such as the production of meat from other animals, except for pigs, poultry, sheep and goats, are of some degree of importance in relation to the dependent variable. The observed pattern can be attributed to a pre-assembled structure in the agro-industrial sense, in which it allows the slaughter of animals of another nature. In general, these aspects must be thoroughly evaluated, in order to obtain more convincing results, especially those that can guide public policies effectively.

COMMITTEE MEMBERS:
Externo à Instituição - ELIAS MENDES COSTA - OUTRO
Presidente - 1548043 - MARCO ROBERTO BOURG DE MELLO
Interno - 2316004 - RONDINELI PAVEZZI BARBERO