Conceptual model for detecting favorable conditions of coffee pests in a smart farming environment

Background: Crop pests are among the greatest threats to food security, generating broad economic, social, and environmental impacts. The impacts of crop pests can be reduced by identifying the conditions that generate them early. These pests interact with their hosts and the environment through complex pathways, and it is increasingly common to find professionals from different areas (farmers, technicians, plant pathologists, computer scientists, economists, sociologists, etc.) gathering into projects that attempt to deal with that complexity most often involving several crop pests. A pest development forecasting can be made using prediction models and it is required for three reasons: economic impact, safety, and justification of control methods. Given this situation, it is necessary to build interdisciplinary work guides that allow the construction of models for the comprehensive management of pest development capable of overcoming the challenges imposed by the presence or absence of data. Aims: Propose a conceptual model for the detection of favorable conditions for coffee pests in a smart farming environment, based on the use of data value and variety, and expert knowledge. Methods: Starting from theoretical references on the realization of mappings and systematic reviews of the literature, the approach proposes a series of steps that lead to a State of Science as a knowledge base for modeling tasks. The modeling tasks are framed in knowledge-based modeling methodologies, as well as data-based modeling. Results: A conceptual model that guides activities for modeling and forecasting the development of diseases and pests in crops, where implementation details are subject to existing methodologies and frameworks. Forecasting solutions can be approached through models based on knowledge and data, according to the requirements and available elements of the person or group of people who will carry out the modeling based on the proposed processes. Additionally, a phase for the exploration of the complementarity between the generated models is proposed. The conceptual model is applied for the development of coffee diseases and pests as a specific case study. Conclusions: Our approach presents a comprehensive conceptual model that guides a robust crop pest modeling process, from obtaining knowledge of the crop pest to be modeled, to the modeling alternatives according to the available resources necessary for modeling such as data and knowledge. For example, a common problem is the amount of data with which the models are trained. If the data is not enough, a modeling alternative that does not require data is needed. Several approaches about crop pest modeling assume knowledge of the problem that is already present, without considering steps to obtain and refine it, and others carry out the modeling process empirically without following a methodology. Although this does not mean that the results are less reliable, the use of methodologies is recommended to achieve an orderly, reliable and well-presented process