Machine Learning-Based Corn Yield Forecasting Using Meteorological and Agronomic Data
Develop a machine learning algorithm able to forecast the expected yield at the end of the season by analising weather data and field management data.
Requirements
- M.Sc. in Machine Learning, Data Science, Computer Science, Mathematics, Telecommunications, or similar
- Knowledge of Python
- Software development skills
- Basic concepts of timeseries processing
- Basic concepts of data science, concerning data analysis, processing and machine learning
Description
Accurately forecasting corn yield is crucial for agricultural planning, food security, and economic decision-making. Traditional yield estimation methods rely on historical trends, statistical models, or complex process-based simulations. However, recent advancements in machine learning (ML) provide an opportunity to enhance yield predictions by leveraging diverse datasets, including meteorological data, field management practices, and crop phenology information. This thesis aims to explore and develop machine learning-based models for corn yield forecasting using a dataset collected internally. The research begins with an extensive review of existing methodologies for corn yield forecasting using machine learning. Once a solid foundation is established, selected methods from the literature will be replicated to assess their performance on the available dataset. Based on this evaluation, the study will introduce novel enhancements to existing models to improve accuracy and robustness. A key focus will be understanding the impact of different data sources, such as weather variables, field management records, phenology stages, and maize variety, on the predictive performance of machine learning models. This research aims to provide a detailed comparative analysis of machine learning methods for corn yield forecasting using diverse datasets. The proposed methodology will incorporate innovative techniques to enhance predictive accuracy and reliability. By assessing the influence of meteorological and agronomic factors, the study will offer insights into how different variables contribute to yield forecasting. The final outcome will be a reproducible framework for data-driven agricultural forecasting, with potential applications in farm management, policymaking, and agribusiness decision-making. Machine learning holds great promise for improving the accuracy and reliability of corn yield forecasting. By integrating diverse data sources and leveraging advanced modeling techniques, this research aims to refine predictive methods, ultimately supporting farmers, policymakers, and agricultural stakeholders in making more informed decisions. The findings of this study have the potential to contribute to sustainable agricultural practices, increased productivity, and more efficient resource allocation in the farming sector.