A hybrid machine learning strategy for aquatic plant surveillance in sustainable aqua-ecosystems using IoT attributes

Abstract

Using a hybridized machine learning framework combined with IoT technology, this research proposes a unique way of monitoring and maintaining the optimal condition of aquatic plants in ecosystems. Our method integrates state-of-the-art convolutional neural networks for plant health data analysis, time series analysis for water quality temporal variation interpretation, and clustering algorithms for ecological data pattern identification. We used an extensive open-source dataset that included various characteristics of IoT platforms, such as water temperature, pH, turbidity, dissolved oxygen levels, and particular metrics for plant development. Because of its critical role in maintaining ecological harmony in our research region, a certain aqua-plant is the primary focus of the dataset. Our Hybrid Machine Learning Strategy (HMLS) demonstrated outstanding performance in forecasting aquatic plant health and growth patterns with a Graph Neural Network (GNN), obtaining a 94 % accuracy rate in plant health data assessments and categorization. In addition, the contour-based clustering technique was employed to effectively group comparable ecological circumstances with a 93.5 % accuracy rate, while time series analysis identified temporal changes with a 94.22 % forecast accuracy. These results prove that our integrated strategy serves a purpose for aqua-ecosystem sustainability management by offering predicted insights and real-time monitoring. This study substantially contributes to environmental monitoring by using modern machine learning algorithms and IoT technologies to provide a scalable and economical solution for aquatic ecosystem conservation. © 2025 Elsevier B.V.