计算机工程与信息技术学报

Machine Learning Based Yield Prediction System

Om Mane*, Chanthrika RL, Tanvir Mungekar, Preethi Sai Yelisetty, Bindhu Sree and Jaya Subalakshmi R

The wild blueberry industry is a significant contributor to the agricultural sector in the northeastern United States. However, unpredictable weather conditions, soil variability, and pest infestations can significantly affect crop yield, leading to losses for farmers and stakeholders. Therefore, the development of precise and credible crop yield prediction models is critical for efficient resource allocation, improved crop management, and effective marketing strategies. Machine learning algorithms, such as decision tree, linear regression, XGBoost, LightGBM, random forest, AdaBoost, histogram gradient boosting, and CatBoost, have shown great potential for crop yield prediction in recent years. These algorithms can analyze large datasets, identify patterns, and create accurate projections, offering farmers with essential insights into the management of crops, future yields, and commercialization. The models developed in this study can enable farmers and stakeholders to make informed decisions about crop planning and resource allocation, improving the efficiency and sustainability of the wild blueberry industry. Additionally, market forecasters can use these models to predict future demand for wild blueberries, aiding in the development of effective marketing strategies. In conclusion, the development of accurate and reliable crop yield prediction models by application of machine learning algorithms holds the prospective to have a big impact on the field of agriculture, particularly for industries such as wild blueberries that are vulnerable to weather variability and other factors.