MLCAS2021 launches Crop Yield Prediction Challenge

In this competition, you will utilize a dataset consisting of 103,365 performance records over 13 years and 150 locations with weekly weather data for each location throughout the growing season of 30 weeks. In particular, you will develop machine learning algorithms that can accurately predict yearly crop yield for a particular performance record.

This challenge is part of the MLCAS2021 workshop. Accurate prediction of crop yield can improve agricultural breeding and provide monitoring across diverse climatic conditions. This can protect crop production from climatic challenges. To predict crop yield, it is important to integrate weather information across the crop growing season for multiple genotypes. Unraveling the importance of different weather parameters for crop yield prediction would be a substantial step forward in understanding the impact of climate change on a variety’s plasticity.

Our dataset assimilates and utilizes complex data for crop yield prediction. This dataset can provide a significant advancement in the domain of yield prediction - previous works have mostly relied on a small subset of factors, therefore failing to capture the complexity of biological interactions and more site-specific weather variable complexities. Some approaches have been based on meteorological data (maximum daily temperature, minimum daily temperature) without field-scale farming data. It is important to develop models which can capture temporal correlations and our dataset can enable that.

Link to website