Identifying key climatic factors for forecasting wheat yield is crucial for developing effective strategies to adapt agricultural practices. This helps mitigate the adverse effects of climate change on wheat production, particularly in areas near Astana, the second coldest capital in the world. The goal of this research is to build machine learning models, such as linear regression, decision trees, and boosting algorithms, to determine the weather variables that most influence wheat yield. The study relies on wheat yield and meteorological data from the Akkol district in the Akmola region. The data include air temperature, humidity, precipitation, wind speed and direction, soil surface temperature, and air humidity deficit.
In this research, six linear machine learning models were implemented to build predictive frameworks, four decision tree-based models and two boosting algorithms were tested. The results revealed that several features significantly impact wheat yield. Decision tree models outperformed others in prediction accuracy. These findings were interpreted and can provide valuable insights for making informed decisions in agricultural management.