Notebook Examples
These interactive notebooks provide in-depth examples of TStrends functionality.
Note
The full notebooks contain many interactive elements and visualizations which may take time to load. Below are links to each notebook with a brief description of what they contain.
Trend Labelling Catalogue
This notebook demonstrates all the available trend labellers and shows their behavior across different parameters.
Content: Examples of Binary CTL, Ternary CTL, Binary Oracle, and Ternary Oracle labellers
Visualizations: Multiple plots showing labelling results with different parameters
Advanced Topics: Parameter sensitivity analysis
How to Use Simple Labellers
Learn how to use the basic trend labelling functionality in TStrends.
Content: Step-by-step guide to using labellers
Examples: Finding good parameters for labellers
Advanced Topics: Return estimation with different fee structures
Parameter Optimization Example
This notebook shows how to optimize trend labeller parameters using Bayesian optimization.
Content: Optimizing Ternary Oracle labeller parameters
Examples: Applying optimization to different time series
Advanced Topics: Handling different time series characteristics
Label Tuning Example
This notebook demonstrates how to transform discrete trend labels into continuous values expressing trend potential.
Content: Using the
RemainingValueTunerto enhance trend labels and optional postprocessor pipelinesExamples: Tuning parameters, smoothing, forward-looking filters, and temporal shifts
Advanced Topics: Visualizing the effect of tuning on prediction models