Tutorials

This section provides detailed, step-by-step tutorials to help you get started with FinMLKit. Each tutorial focuses on a specific aspect of the library, guiding you through its functionalities with examples and explanations.

• Why FinMLKit?
• The problem we’re tackling
  • 1. Time-based sampling bias
  • 2. Inadequate labeling
  • 3. Performance bottlenecks
• What FinMLKit brings
  • Three principles
  • Concrete outcomes
• How this advances research
• What’s different from existing libraries
• Open source philosophy
• Call to action
• Processing Raw Trade Data
  • Downloading Raw Trade Data
  • Preprocessing the Data
  • Key Features of TradesData
  • Next Steps
• Saving and Loading Data
  • Saving Data
  • Inspecting the HDF5 File
  • Adding Time Bars
  • Reading Time Bars
  • Next Steps
• Building Intra-Bar Features
  • Restoring Preprocessed Data
  • Building Time Bars
  • Directional Features
  • Size Distribution Features
  • Footprint Features
  • Next Steps
• Building Inter-Bar Features
  • Defining Transforms
  • Custom Transforms
  • Building Feature Kits
  • Next Steps
• Feature pipelines: Compose, FeatureKit and the computation graph
  • Prerequisites
  • Dataset
  • Chaining transforms with Compose
  • Batch execution and caching with FeatureKit
  • Visualizing dependencies with ComputationGraph
  • Reproducibility: save and load pipeline configurations
  • Tips
  • Integrating external libraries (e.g. TA-Lib) with ExternalFunction
• Building Labels
  • Setting Up Target Returns
  • Building TBM Labels
  • Sample Weights
  • Final Weights
  • Next Steps