Assessment of Marine Ecosystem Health Based on Transformer Architecture

High School Attached To Shandong Normal University, Jinan City, 250014, China

^* Corresponding author’s e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.

Abstract

Aiming at the key scientific problem of marine ecosystem health assessment, an innovative intelligent assessment model based on Transformer architecture and Particle Swarm Optimization (PSO) algorithm is proposed in this paper. By integrating satellite remote sensing, buoy monitoring and other heterogeneous ocean data, a complete assessment framework including data preprocessing, feature extraction and state prediction is constructed. PSO algorithm is used to optimize the hyperparameters of Transformer model systematically, which significantly improves the ability of the model to capture complex spatiotemporal characteristics of marine ecosystem. The experimental results show that the RMSE of PSO-Transformer model is reduced to 0.87, and the coefficient of determination (R²) is 0.941. The performance of PSO-Transformer model is significantly better than traditional ARIMA, SVR and standard Transformer model without optimization. This paper provides a new technical path for marine ecosystem health assessment. This paper explores the application of the self-attention mechanism in the multi-parameter fusion analysis of marine ecosystems and develops an optimization method for deep-learning models based on swarm intelligence.