ABSTRACT
This study is dedicated to the development of a methodology based on supervised machine learning for soybean classification and justified as technological innovation to predict whether soybean classification is in the standard or non-standard established by normative instruction No. 11/2007 of the Ministry of Agriculture, Livestock, and Food Supply (MAPA). This study aimed to develop a website using supervised machine learning to classify soybeans, providing an assertive decision-making process in real-time. A technological tool was created to assist the farmer and the storage unit in the classification of soybeans, considering the perceived reality and potential instruments consistent with the reality of the area. Therefore, a website in Python language was created using the Pandas, Pandas Profiling, Seaborn, Matplotlib, NumPy, Scikit-learn, PyCaret, and Streamlit libraries. In the end, the system could predict whether the soybean is in the standard or non-standard established by the soybean classification normative. In this sense, the results showed the robustness and precision of the proposed new methodology.
KEYWORDS
standard; non-standard; decision-making