Implementation of models to predict avocado exports

Abstract

Business intelligence (BI) integrates and analyzes the behavior of historical data streams, obtaining predictions. The project allows companies in the agricultural sector dedicated to the cultivation and export of avocados to make decisions based on artificial intelligence, promoting growth and competitiveness in the market. Models that apply data analytics are implemented through simple linear regression and recurrent neural networks (RNN). To carry out the project, data extraction, transformation, and loading (ETL) processes were used. The coding was developed in Python with the Django framework, using the sklearn, linear_model, LinearRegression, seaborn, stastsmodels and tensorflow libraries, among others. In the predictions for the years 2016 to 2021, greater precision was verified in the linear regression model. When making the export projection for the next six years, the results coincide with a minimal difference between the two models.