Anomaly Detection on Univariate Sensing Time Series Data for Smart Aquaculture Using Deep Learning

Abstract Aquaculture plays a significant role in both economic development and food production. Maintaining an ecological environment with good water quality is essential to ensure the production efficiency of aquaculture. Effective management can prevent abnormal conditions contribute significantly security. Detecting anomalies aquaculture crucial that maintained correctly meet healthy proper requirements for fish farming. This article focuses on use deep learning techniques detect data environment. Four anomaly detection techniques, including Autoencoder, Variational Long-Short Term Memory Spectral-Residual Convolutional Neural Network, were analysed using multiple real-world sensor datasets collected from IoT systems. Extensive experiments conducted temperature, dissolved oxygen, pH parameters, evaluation analysis revealed Autoencoder method showed promising results detecting temperature oxygen datasets, while Network demonstrated best performance datasets.