A bibliometric analysis of the literature on crop yield prediction: insights from previous findings and prospects for future research.

Abstract

This research presents a bibliometric analysis of articles predicting crop yield using machine learning methods. While several systematic review articles exist, a comprehensive bibliometric analysis illustrating the knowledge structure and research trends, along with collaboration networks among authors, institutions, and countries in this field, has not been conducted. The study focused on 826 articles published over a 32-year period (1992 to 2023) and revealed a significant increase in publications, particularly in recent years. Zhang Zhao from China authored the majority of articles, while the highest number of citations was associated with articles by Zhu Yan and Lobell. Leading countries in article publications are the USA, China, India, Germany, Australia, and Canada, showing strong interconnections in citing each other's research. The Chinese Academy of Sciences and the US Department of Agriculture are the institutions with the highest number of articles and citations in this domain. The journals Agricultural and Forest Meteorology and Remote Sensing are recognized as top ranking journals in this field (Q1). Based on co-occurrence analysis, three main thematic domains were identified: weather and crop yield prediction, plant growth simulation models, and crop yield prediction using remote sensing data. The research suggests a focus on variables such as disease, pests, insects, and soil salinity when predicting yield. Additionally, greater attention should be given to discussions on food security and crop yield, especially in developing countries.