Inside the LSTM: An XAI Field Guide to Weather Prediction

Short summary

A hands-on tutorial for training an LSTM on daily temperature data and then applying three explainability methods—permutation importance, SHAP, and Integrated Gradients—to understand which input timesteps drive predictions. Covers data windowing, scaling, model architecture with stacked LSTM layers, and early stopping. Each XAI method is implemented in code with practical interpretation of results.

• •Train a stacked LSTM on 7-day temperature windows to predict day 8
• •Apply permutation importance, SHAP, and Integrated Gradients for interpretability
• •Includes full Keras code for model building, training, and XAI analysis

Generated with AI, which can make mistakes.