Análisis de supervivencia de la producción del camarón segmentado por genética y alimentación.

Abstract

The research conducted, entitled “Survival Analysis of Shrimp Segmented by Genetics and Feeding,” aims to analyze the probability of shrimp survival in the Ecuadorian productive sector by examining key production variables such as genetics and type of feeding. This research arises from the need to develop a tool that makes it possible to determine the best suppliers who, based on historical data, enable specific genetic lines to reach a certain weight under a defined feeding scheme, as well as to identify their optimal harvest time. Methodologically, the study focused on a quantitative approach based on supervised models, specifically logistic regression, in order to estimate the probability of shrimp survival using historical production data. To validate the results, performance metrics such as the confusion matrix, accuracy, precision, recall, F1 score, as well as the ROC curve and AUC were used, allowing the model’s performance to be properly evaluated. Additionally, production indicators were integrated to complement the technical analysis. The findings show that survival does not depend exclusively on feeding techniques, but rather on the combination of genetics and the supplier responsible for developing those genetic lines. This analysis highlights the need to evaluate and classify both suppliers and the genetic lines that have been used.