Research Trends in Machine Learning Applications for Predicting Ecosystem Responses to Environmental Changes

Abstract

This research discusses the trends in machine learning (ML) applications for predicting ecosystem responses to environmental changes. A keyword search was conducted in the WoS database using Boolean operators to identify relevant peer-reviewed articles. The search focused on English-language documents published between 2014 and 2023, while excluding non-original articles. Bibliometric data, includingpublication trends, citation counts, author collaboration patterns, and keyword analysis, were extracted from 554 retrieved articles. The data was then analyzed and visualized using R and VOSViewer. The study highlights the significant growth in annual scientific production, reflecting a growing interest in thisinterdisciplinary field. Core concepts such as “climate change,” “biodiversity,” and “ecological responses” continue to receive significant attention, while contemporary themes like “variability,” “time-seriesanalysis,” and “organic matter” are emerging. Co-authorship networks demonstrate extensive collaborationsacross countries, with the United States and China playing prominent roles. The research topics have evolvedfrom “ecological responses” and “community” to a focus on “model,” “optimization,” and “performance,” with an emphasis on fine-tuning models to incorporate climate variability.