Scholarlink

<p>Artificial intelligence (AI) is rapidly transforming aquaculture by enhancing productivity, environmental monitoring, and decision-making. However, the scientific literature on AI applications in this field remains fragmented, with limited understanding of its thematic structure, geographic concentration, and emerging trends. Addressing this gap is crucial for guiding research, investment, and policy efforts toward sustainable innovation in aquaculture. This study adopts a quantitative, exploratory, and descriptive design to map the global scientific production on AI in aquaculture. A total of 2610 documents published between 1981 and 2025 were extracted from Scopus and Web of Science. The analysis integrates bibliometric indicators, unsupervised topic modeling using Latent Dirichlet Allocation (LDA), and multivariate visualization through multidimensional scaling (MDS). The results reveal a sustained growth in publications, with an annual rate of 13.14%. China, India, and the USA dominate in output, yet international collaboration remains low (5.38%). LDA identified 25 latent topics grouped into four macro-areas: intelligent sensing and automation, fish health and genomics, environmental monitoring, and computer vision. Temporal analysis shows a shift toward real-time water quality prediction, YOLO-based object detection, and AI-driven disease diagnosis. However, topics such as decision support systems, integrated multi-trophic aquaculture (IMTA), and explainable artificial intelligence (XAI) remain underexplored. Visualization techniques revealed patterns of thematic clustering and highlighted geographical disparities in topic focus and adoption. These findings suggest that AI research in aquaculture tends to be concentrated on technological optimization, with comparatively limited representation of social equity, ethical frameworks, or inclusive innovation. The Global South remains underrepresented, both in scientific production and in the contextual adaptation of AI tools. This study provides an exploratory mapping of the research landscape and offers guidance for promoting responsible, sustainable, and regionally relevant AI integration in aquaculture.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Exploring the scientific landscape of artificial intelligence in aquaculture: trend and topic analysis using unsupervised machine learning and multivariate visualization

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月1日

查看全文
  • Javier De La Hoz-M,
  • Sara Cruz-Botto

摘要

Artificial intelligence (AI) is rapidly transforming aquaculture by enhancing productivity, environmental monitoring, and decision-making. However, the scientific literature on AI applications in this field remains fragmented, with limited understanding of its thematic structure, geographic concentration, and emerging trends. Addressing this gap is crucial for guiding research, investment, and policy efforts toward sustainable innovation in aquaculture. This study adopts a quantitative, exploratory, and descriptive design to map the global scientific production on AI in aquaculture. A total of 2610 documents published between 1981 and 2025 were extracted from Scopus and Web of Science. The analysis integrates bibliometric indicators, unsupervised topic modeling using Latent Dirichlet Allocation (LDA), and multivariate visualization through multidimensional scaling (MDS). The results reveal a sustained growth in publications, with an annual rate of 13.14%. China, India, and the USA dominate in output, yet international collaboration remains low (5.38%). LDA identified 25 latent topics grouped into four macro-areas: intelligent sensing and automation, fish health and genomics, environmental monitoring, and computer vision. Temporal analysis shows a shift toward real-time water quality prediction, YOLO-based object detection, and AI-driven disease diagnosis. However, topics such as decision support systems, integrated multi-trophic aquaculture (IMTA), and explainable artificial intelligence (XAI) remain underexplored. Visualization techniques revealed patterns of thematic clustering and highlighted geographical disparities in topic focus and adoption. These findings suggest that AI research in aquaculture tends to be concentrated on technological optimization, with comparatively limited representation of social equity, ethical frameworks, or inclusive innovation. The Global South remains underrepresented, both in scientific production and in the contextual adaptation of AI tools. This study provides an exploratory mapping of the research landscape and offers guidance for promoting responsible, sustainable, and regionally relevant AI integration in aquaculture.