Data-Driven Ecosystem Modeling for Sustainable Fish Species Management in Protected Areas Using the FP-Growth Algorithm

This study presents a data-driven framework for supporting conservation planning through the extraction of actionable ecological insights using association rule mining (ARM). Focusing on 13 fish species within Romania's Natura 2000 protected areas, we apply the FP-Growth algorithm to ecological and management datasets to uncover frequent co-occurrence patterns among conservation actions. By encoding expert-evaluated habitat indicators into binary transactional data, we identify 44 high-confidence association rules that reveal interdependent management measures crucial for maintaining favorable species conservation statuses. These rules provide evidence-based guidance for designing integrated, synergistic conservation strategies, enhancing decision-making efficiency for stakeholders and policymakers. Our approach demonstrates the utility of ARM in ecological modeling, offering a scalable and replicable method for adaptive management in complex protected area networks. Furthermore, by aligning our findings with foundational ecological theories such as ecosystem-based management and metapopulation dynamics, we enhance the ecological interpretability and theoretical grounding of the extracted rules.