Probability on Graphical Structure: A Knowledge-Based Agricultural Case

Autor: Ianishi, Paula, Gonzatto Junior, Oilson Alberto, Henriques, Marcos Jardel, Nascimento, Diego Carvalho do, Mattar, Gabriel Kamada, Ramos, Pedro Luiz, Ara, Anderson, Louzada, Francisco
Zdroj: Annals of Data Science; April 2022, Vol. 9 Issue: 2 p327-345, 19p
Abstrakt: This paper provides a rich framework to estimate the causal relationship among eighteen features (related to the product type and classification) on an agronomy study by using Bayesian Networks, which are a type of probabilistic graphical model. Thereby, with this class of models, we aimed to classify and identify the complaints based on corn seed commercialization. Simulation studies were used to compare both adopted algorithms, K2 and PC, and their hybrid version. These studies indicate excellent classification performance, given the knowledge of the network structure. After the estimated Directed Acyclic Graph, three features (Brand, Germination percentage, and Amount of commercialized bags) were evidenced as Impacting factors in the complaints based on corn seed commercialization.
Databáze: Supplemental Index
Externí odkaz: