Central and western Baltic Sea data and ecosystem risk assessment for 2020

Open data repository for central and western Baltic Sea data and results from two ecosystem risk assessment models for 2020. The models were a Multi K-means-based unsupervised risk assessment process and a deep learning-based model (Variational Autoencoder, VAE) with the consecutive re-application of Multi K-means to the anomalies identified by the VAE.本开放数据仓库收录波罗的海中部与西部海域的相关数据，以及2020年两款生态系统风险评估模型的输出结果。两款模型分别为基于多K均值聚类（Multi K-means）的无监督风险评估流程，与基于深度学习的变分自编码器（Variational Autoencoder, VAE）模型：后者将多K均值聚类算法依次复用于该变分自编码器所识别的异常样本。