DeepDow

deepdow (read as “wow”) is a Python package connecting portfolio optimization and deep learning.

USING DEEPDOW:

• Installation
  • Development
• Introduction
  • Name
  • Traditional portfolio optimization
  • Why DeepDow different?
  • References
• Basics
  • Data
  • Predictions AKA weights
  • Loss
  • Assumptions
• Data Loading
  • Introduction
  • Raw data
  • raw_to_Xy
  • InRAMDataset
  • Dataloaders
• Benchmarks
  • Benchmark class
  • Simple benchmarks
• Layers
  • Introduction
  • Transform layers
  • Collapse layers
  • Allocation layers
  • Misc layers
  • References
• Networks
  • Existing networks
  • Writing custom networks
• Losses
  • Introduction
  • Definitions
  • Portfolio returns
  • Available losses
  • Arithmetic operations
• Experiments
  • History
  • Callbacks
• Examples
  • End to end
  • Layers

DEVELOPMENT

• Changelog
  • v0.2.1
  • v0.2.0

API Reference:

• deepdow package
  • deepdow.benchmarks module
  • deepdow.callbacks module
  • deepdow.data module
  • deepdow.experiments module
  • deepdow.explain module
  • deepdow.layers package
  • deepdow.losses module
  • deepdow.nn module
  • deepdow.utils module