<![CDATA[This study proposes a microservice-oriented machine learning framework to enhance intelligence and scalability in perishable fish cold chain logistics. Unlike conventional monitoring-centric systems, the framework integrates edge–cloud computing with multimodal machine learning models, including random forest for anomaly detection, long short-term memory (LSTM) for spoilage risk prediction, and convolutional neural network (CNN) for visual fish quality classification. The research adopts a design science approach combining literature analysis, field observations at cold storage facilities in Indramayu, Indonesia, and simulation-based validation. Experimental results demonstrate the feasibility of distributed analytics, modular deployment, and real-time inference within heterogeneous logistics environments. The proposed framework provides a deployable architectural reference for intelligent fisheries cold chain management and supports future large-scale, multi-stakeholder implementation.]]>
Copyrights © 2026
