Diseño de un modelo predictivo para la oferta de los productos agrícolas del municipio de Simijaca Cundinamarca

Abstract

The project aims to design a predictive model of the supply of the main agricultural products grown in the municipality of Simijaca, allowing farmers in the area to make better decisions based on projections projected in a given period, thus reducing the risk and uncertainty of investments. In order to achieve the proposed objective, an initial recognition of the main agricultural products grown in the municipality is carried out by reviewing the production history and evaluating 80-20, and then determining the different variables that affect the entire production process of each of products from the analysis of the agricultural computer bulletins published by DANE in the last ten years. Subsequently, the different forecast models studied in the theoretical framework were evaluated with historical data on the supply of corn, potatoes, beans, peas, bulb onions and carrots. With the results found, it was determined that an approximation to the prediction of supply of agricultural products can be made using two models, the first constitutes a multiple linear regression which groups half of the products and the rest with the moving average method. The predictions produced by these models allow us to conclude that supply can be modeled with these techniques despite the fact that the indicators are not the most acceptable.