FramePests: A Comprehensive Framework for Crop Pests Modeling and Forecasting

Crop pests are among the greatest threats to food security, generating broad economic, social, and environmental impacts. These interact with their hosts environment through complex pathways, it is increasingly common find professionals from different areas gathering into projects that attempt deal this complexity. We propose a framework called FramePests guiding steps activities for crop pest modeling forecasting. From theoretical references about carrying out mappings systematic reviews of literature, proposes series leading state science as knowledge base tasks. Then, two solutions, based on data used. Finally, model outputs performances compared. The application proposed was demonstrated coffee leaf rust modeling, which we obtained data-based built using gradient boosting algorithm ( xmlns:xlink="http://www.w3.org/1999/xlink">XGBoost ) mean absolute error 7.19% knowledge-based represented by hierarchical multi-criteria decision structure an accuracy 56.03%. A complementary study our case allowed us explore how elements can improve model, improving its 7.07%. showed be alternative when available dataset has approximately 60 instances. Data-based models tend have better performance, but replicability conditioned diversity in Knowledge-based may simpler allow expert supervision, these not usually tied specific sites.