Diseño de un modelo predictivo de la oferta de aguacate hass en el Municipio de Herveo Tolima

Abstract

The research project aims to design a predictive model to recognize the variables that affect the production of the agricultural sector in relation to the hass avocado type grown in the municipality of Herveo, Department of Tolima, so that farmers in this area access this information, facilitating decision-making and making them more adjusted , but above all taking into account the forecasts projected over the defined time period, thus contributing to reducing and reducing the risk in the production of avocado cultivation of this particular type.

To achieve the proposed objective, the variables were identified and analyzed with historical data collected from various sources such as the production, yield area of harvesting and production of Hass type avocado at the national, departmental and municipal level, determining their correlation with the production, supply and demand of the avocado, which allows to establish the level of supply of the avocado in the municipality of Herveo, based on a medium-term projection horizon.

Afterwards, an evaluation of several forecast models studied in the theoretical framework was made with historical data in relation to the municipality of Herveo, such as multiple linear regression, the time series model, the moving average model and the Simple Exponential Smoothing model where the production and harvested area variables were evaluated, the forecast of the growth of avocado production.

From the results found it was established that you can make through the time series forecast model a prediction of avocado crop supply, specifically the Hass type. In addition, concluding that the production of Hass avocado over the period of time analyzed has linear behavior, although the predicted production is less than actually recorded.