Application of machine learning algorithms for predicting agricultural crop yields

Abstract

This article examines the use of machine learning algorithms for predicting the yield of agricultural crops. The primary classification method chosen is the C4.5 algorithm, which allows for the construction of interpretable models that identify key factors affecting yield. The analysis utilized data from a dataset available on the Kaggle platform, including information on various crops, their yields, and associated factors such as rainfall, fertilizer usage, air temperature, and the content of nitrogen, phosphorus, and potassium in the soil. The conducted correlation analysis showed that air temperature and the content of nitrogen, phosphorus, and potassium in the soil have the greatest impact on yield. Despite high correlation, the amount of fertilizer and rainfall were less significant in the model, indicating the need for further investigation of their influence. The model evaluation on the Deductor Studio platform demonstrated high classification accuracy, but there are opportunities for improvement. The importance of the results underscores the necessity for precise monitoring and management of key factors in agricultural practice to enhance productivity. Future research could focus on integrating larger datasets and more complex algorithms, as well as utilizing Internet of Things (IoT) systems for more accurate monitoring and yield prediction.