Análisis histórico y aplicación de modelos predictivos del precio de la tilapia Ecuatoriana

Abstract

Tilapia is an aquaculture product that once represented a strong income for Ecuador, however, today it does not represent even half of what it once was. Therefore, this project aims to develop univariate econometric models to predict the price of Tilapia exported to the United States by the year 2025. This information will allow Ecuadorian exporters to generate strategies to become more competitive in the international market and increase their profits. In addition, we also analyzed the historical behavior of the price and compared it with the average price of tilapia in the Latin American market. We obtained 272 observations for each variable, with monthly periodicity data. The models that were tested are ARIMA, SARIMA and Random Forest, for the application of these models it was necessary to perform statistical analysis and tests and in certain occasions it was necessary to induce stationarity. As a result, it was observed that the Ecuadorian series showed a downward trend and a seasonal pattern, so the ideal model was SARIMA. For the Latin American data, it was noted that the data circulated around a mean and high volatility in prices, it was determined that the best fit model was ARIMA. Thanks to the SARIMA model, we obtained a prediction consistent with the data and reliable so that exporters know in advance how prices will behave in the market, anticipate price peaks and troughs and can generate strategies to cope with any scenario that may arise. Keywords: Tilapia, Export Prices, Price Forecasting, Econometric Models.