ARIMA: UN MODELO PREDICTIVO PARA LA ESTIMACIÓN DE LA PRODUCCIÓN AGRÍCOLA ECUATORIANA

Abstract

Database management is an important tool for decision-making if it is complemented with mathematical tools that allow their modeling with the purpose of predicting the behavior of the data in the future and taking data for planning, particularly in a sector with as much uncertainty as agriculture. One of the models that have been used with the greatest success is ARIMA, which is based on the interpretation of statistical data through regression analysis, allowing predictions to be made with great efficiency. To analyze the scope of its applications in the agricultural field, a systematic review was made using the PRISMA methodology by reviewing 180 articles in scientific databases such as Scopus, Scielo, Latindex, Redalyc and google evidence, of which 24 articles were selected. According to the inclusion criteria used in the search, the results show that the ARIMA model has been successfully used to estimate yields and prices in the production of cereals such as corn and cinchona and in crops of great importance for the Ecuador as it is in cocoa and coffee, where the estimation of yields and price variations is vital due to its role in foreign trade. Despite the fact that new models based on fuzzy logic, neural networks and artificial intelligence have emerged, the ARIMA model turns out to be efficient when compared with these models, with the advantage that its management and implementation is easy for most users.