Skip to content

Latest commit

 

History

History
112 lines (91 loc) · 3.44 KB

CHANGELOG.md

File metadata and controls

112 lines (91 loc) · 3.44 KB

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[1.0.0] - 2025-10-14

Added

  • Model Architectures

    • LSTM Baseline implementation
    • MC Dropout LSTM for uncertainty quantification
    • Bayesian Neural Network (Pyro-based)
    • Transformer architecture for time series

  • Data Pipeline

    • Yahoo Finance data fetching (scripts/fetch_data.py)
    • Technical indicators calculation (21 indicators)
    • Data preprocessing and normalization
    • Train/validation/test splitting

  • Training Infrastructure

    • Modular training loop with early stopping
    • Gradient clipping for RNN stability
    • Model checkpointing
    • Configuration-based experiments

  • Evaluation & Analysis

    • Comprehensive metrics (RMSE, MAE, R², MAPE)
    • Backtesting system with strategy evaluation
    • Uncertainty-aware position sizing
    • Performance visualization

  • Hyperparameter Optimization

    • Optuna integration for HPO
    • TPE algorithm for efficient search
    • 5-parameter search space (hidden_size, num_layers, dropout, lr, batch_size)

  • Documentation

    • Comprehensive README.md
    • PROJECT_ANALYSIS_REPORT.md (detailed technical analysis)
    • QUICK_REFERENCE.md (at-a-glance guide)
    • FIGURES_DOCUMENTATION.md (all visualizations documented)
    • HPO_SEARCH_SPACE.md (hyperparameter details)
    • TECHNICAL_INDICATORS_TABLE.md (all 21 indicators with formulas)
    • CONTRIBUTING.md (contribution guidelines)

  • Visualizations

    • Figure 6: AAPL Price History (2015-2024)
    • Figure 7: Feature Correlation Heatmap
    • Figure 8: Training/Validation Loss Curves
    • Figure 9: Uncertainty Bands (COVID-19 demonstration)
    • Figure 10: Cumulative Returns Comparison
    • Professional table images for reports

  • Web Application

    • Streamlit interactive dashboard
    • Model prediction interface
    • Visualization panels
    • Configuration options

  • Testing

    • Unit tests for models (test_models.py)
    • Unit tests for preprocessing (test_preprocess.py)
    • Unit tests for metrics (test_metrics.py)
    • Project launcher with test execution

  • Utilities

    • Figure generation scripts
    • Table image generation
    • Configuration management
    • Results logging

Project Organization

  • Organized code into professional structure
  • Created utils/ for utility scripts
  • Established reports/ for documentation
  • Set up configs/ for YAML configurations
  • Implemented src/ modular architecture

Configuration Files

  • configs/lstm_baseline.yaml
  • configs/mc_dropout_lstm.yaml
  • configs/bnn_vi.yaml
  • configs/transformer_baseline.yaml
  • configs/app.yaml

Scripts

  • train.py - Main training script
  • evaluate.py - Model evaluation
  • backtest.py - Strategy backtesting
  • hparam_search.py - Hyperparameter optimization
  • run_project.py - Project launcher (tests + app)
  • utils/generate_figures.py - Documentation figures
  • utils/generate_hpo_table.py - HPO table image
  • utils/generate_indicators_table.py - Indicators table image

[0.1.0] - Initial Development

Added

  • Basic project structure
  • Initial model implementations
  • Data fetching capabilities
  • Simple training scripts

Version History

  • v1.0.0 (2025-10-14): Production-ready release with comprehensive documentation
  • v0.1.0 (Development): Initial implementation phase