Persea americana, commonly known as avocado, is becoming increasingly important in global agriculture. There are dozens of avocado varieties, but more than 85% of the avocados harvested and sold in the world are of the Hass one. Furthermore, information on the market of agricultural products is valuable for decision-making; this has made researchers try to determine the behavior of the avocado market, based on data that might affect it one way or another. In this paper, a machine learning approach for estimating the number of units sold monthly and the total sales of Hass avocados in several cities in the United States, using weather data and historical sales records, is presented. For that purpose, four algorithms were evaluated: Linear Regression, Multilayer Perceptron, Support Vector Machine for Regression and Multivariate Regression Prediction Model. The last two showed the best accuracy, with a correlation coefficient of 0.995 and 0.996, and a Relative Absolute Error of 7.971 and 7.812, respectively. Using the Multivariate Regression Prediction Model, an application that allows avocado producers and sellers to plan sales through the estimation of the profits in dollars and the number of avocados that could be sold in the United States was created.
Crop pests are among the greatest threats to food security, generating broad economic, social, and environmental impacts. These pests interact with their hosts and the environment through complex pathways, and it is increasingly common to find professionals from different areas gathering into projects that attempt to deal with this complexity. We propose a framework called FramePests guiding steps and activities for crop pest modeling and forecasting. From theoretical references about carrying out mappings and systematic reviews of the literature, the framework proposes a series of steps leading to a state of science as a knowledge base for modeling tasks. Then, two modeling solutions, based on data and knowledge are used. Finally, the model outputs and performances are compared. The application of the proposed framework was demonstrated for coffee leaf rust modeling, for which we obtained a data-based model built using a gradient boosting algorithm ( XGBoost ) with a mean absolute error of 7.19% and a knowledge-based model represented by a hierarchical multi-criteria decision structure with an accuracy of 56.03%. A complementary study for our case study allowed us to explore how elements of a data-based model can improve a knowledge-based model, improving its accuracy by 7.07%. and showed that knowledge-based modeling can be an alternative to data-based modeling when the available dataset has approximately 60 instances. Data-based models tend to have better performance, but their replicability is conditioned by the diversity in the dataset used. Knowledge-based models may be simpler but allow expert supervision, and these models are not usually tied to specific sites.