We developed sophisticated AI prediction models using Dynamic Slow Feature Analysis (DSFA), LSTM networks, and Convolutional Block Attention Modules (CBAM) to effectively forecast indoor PM2.5 levels in subway environments.
