Scholarlink

<p>Aquaculture is a vital contributor to global food security and economic sustainability, yet disease outbreaks remain a persistent challenge, causing significant financial losses and threatening fish production. Early and accurate disease detection is critical for effective prevention and treatment, as delayed intervention leads to widespread infections and increased mortality. However, conventional approaches rely on either image-based disease detection or water quality analysis in isolation, often failing to capture the complex interactions between environmental factors and disease occurrence. To address these limitations, this study proposes an integrated approach that combines deep learning-based disease detection with statistical modelling to improve diagnostic accuracy and risk assessment. This research develops a stacked ensemble learning framework for fish health management, utilising a deep learning-based detection model to identify fish diseases and logistic regression to evaluate the impact of pH levels and temperature on disease probability. These models are further refined through a machine learning-based meta-classifier to enhance risk prediction and generate actionable recommendations. One hundred juvenile Red Malaysian Mahseer were analysed, incorporating both fish disease images and water quality parameters to assess system performance. The proposed approach achieved an 87.7% mean average precision for disease identification, ensuring reliable visual-based detection, while the integrated model attained an 85% accuracy in predicting disease risk and recommending mitigation strategies, demonstrating its effectiveness for aquaculture applications. By bridging the gap between image-based diagnosis and water quality analysis, this study offers a comprehensive, real-time diagnostic system that enhances disease management in aquaculture. The findings provide valuable insights for sustainable fish farming, supporting proactive intervention strategies to minimise disease outbreaks and economic losses.</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

Smart aquaculture: an advanced intelligent predictive analysis of disease risks and recommendation system for managing fish health

  • 期刊论文

  • 发表日期：

    2025年5月12日

查看全文
  • M. L. Yasruddin,
  • Z. Husin,
  • M. A. H. Ismail,
  • T. W. Keong

摘要

Aquaculture is a vital contributor to global food security and economic sustainability, yet disease outbreaks remain a persistent challenge, causing significant financial losses and threatening fish production. Early and accurate disease detection is critical for effective prevention and treatment, as delayed intervention leads to widespread infections and increased mortality. However, conventional approaches rely on either image-based disease detection or water quality analysis in isolation, often failing to capture the complex interactions between environmental factors and disease occurrence. To address these limitations, this study proposes an integrated approach that combines deep learning-based disease detection with statistical modelling to improve diagnostic accuracy and risk assessment. This research develops a stacked ensemble learning framework for fish health management, utilising a deep learning-based detection model to identify fish diseases and logistic regression to evaluate the impact of pH levels and temperature on disease probability. These models are further refined through a machine learning-based meta-classifier to enhance risk prediction and generate actionable recommendations. One hundred juvenile Red Malaysian Mahseer were analysed, incorporating both fish disease images and water quality parameters to assess system performance. The proposed approach achieved an 87.7% mean average precision for disease identification, ensuring reliable visual-based detection, while the integrated model attained an 85% accuracy in predicting disease risk and recommending mitigation strategies, demonstrating its effectiveness for aquaculture applications. By bridging the gap between image-based diagnosis and water quality analysis, this study offers a comprehensive, real-time diagnostic system that enhances disease management in aquaculture. The findings provide valuable insights for sustainable fish farming, supporting proactive intervention strategies to minimise disease outbreaks and economic losses.