Further Reading and Resources
Verified core resources to go deeper into what we covered:
- Official PyTorch documentation and tutorials:
- PyTorch Tutorials — Learn the Basics:
Tensor,Dataset,DataLoader,nn.Moduleand the training loop. - PyTorch
nn.LSTMreference: input and output tensor shapes (batch_first=True) and the gate equations. - PyTorch
nn.BCEWithLogitsLossreference: numerical stability and the built-insigmoidin binary classification.
- PyTorch Tutorials — Learn the Basics:
- Time-series validation (walk-forward) and data leakage:
- Scikit-Learn — TimeSeriesSplit: the official explanation of why standard k-fold cross-validation does not work for time series, and of expanding-window (walk-forward) validation.
- Scikit-Learn — Common Pitfalls and Recommended Practices (Data Leakage): how to avoid leakage between training and test data in preprocessing steps such as
StandardScaler. - Hyndman & Athanasopoulos — Forecasting: Principles and Practice (Evaluating Forecast Accuracy): the standard reference on naive (random-walk) baselines and time-series error metrics.
Final reminder: This workshop and guide are for education only and are not investment advice.