Changelog

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

v2023.8.0 (2023-12-13)

Feature

Fix

  • Use validity in rolling fit (087939b)

Documentation

  • Update numpy backend engine notebook (5037603)

v2023.7.2 (2023-09-26)

Fix

  • Convert y into numpy engine array (e92fc34)

v2023.7.1 (2023-09-26)

Feature

  • Support setting the backend engine with environment variable (a40ee5c)

v2023.7.0 (2023-09-26)

Feature

  • Support pytorch in engine (1d9eb07)

v2023.6.5 (2023-09-22)

Fix

  • Use engine linalg in WLS (98ea2ed)

Documentation

v2023.6.4 (2023-08-19)

Fix

  • Revert forcing numpy array conversion (6676fef)

v2023.6.3 (2023-08-17)

Fix

  • Ensure to_numpy converts with the numpy engine array (cd014ce)

v2023.6.2 (2023-08-14)

Fix

  • Remove the type check to support multiple engines (dd0d8a8)

v2023.6.1 (2023-07-26)

v2023.6.0 (2023-07-26)

Feature

  • Relax pandas dependency constraint (1bae2fb)

  • Support various types of numpy engine (05eb58c)

Documentation

v2023.5.1 (2023-06-23)

Feature

  • Add covariance shrinkage (fa59413)

Fix

  • Insufficient returns in fitting rolling risk model (96ca77b)

v2023.5.0 (2023-06-22)

Feature

  • Add covariance estimator (4b5d5d4)

  • Allow passing dict of covariances in accuracy functions (3f2452d)

Documentation

  • Remove unused imports in example notebook (d415101)

  • Update example notebook (cd71b0f)

  • Correct typo (e711966)

  • Update notebook (f9a72a9)

  • Update README and notebook example (cc53ae3)

  • Update crypto statistical risk model notebook (c66f934)

v2023.4.0 (2023-06-05)

Feature

  • Add API to download crypto sample data (6041ada)

  • Support writing and reading from directory (2553366)

  • Allow setting validity in fitting and transforming rolling risk model (32a1f6a)

  • Support asymptotic principal components (APCA) (7169189)

Fix

  • Align parameter n_components type in apca with scikit-learn (3394dbf)

Documentation

  • Update example notebook (b102c02)

  • Add apca section (7f41707)

  • Change the weights to market cap sqrt (20574e4)

v2023.3.0 (2023-03-10)

Feature

  • Support WLS in PCA statistical risk model (71fc3e2)

v2023.2.0 (2023-02-27)

Documentation

  • Correct example configuration (00e136d)

v2023.1.0 (2023-02-12)

Feature

  • Introduce cov half life (92d96dc)

  • Add cov halflife into the accuracy functions (cd20be4)

  • Support exponential weighted factor covariance (ffac512)

  • Support exponential weighted least square (db4ae53)

Fix

Documentation

v2023.0.0 (2023-01-12)

Feature

  • Add value at risk accuracy (7034a8e)

Fix

  • Another bug created from the previous bugfix (d48381a)

  • Bug in verifying ndarry existence (f42bd2d)

  • Typo in parameter (f5c5dab)

  • Correct the forecast return computation (6603ce0)

Documentation

  • Add rolling risk model description (e20bb28)

  • Update README (a8fd716)

  • Correct typo (d15e684)

  • Update bias and VaR documentation page (e9c7542)

  • Add bias and var placeholder pages (55d79b4)

  • Correct factor risk model title (239c41f)

  • Add factor risk model doc (b35ca51)

  • Add risk model descrption (40af5a6)

  • Update README (5877781)

  • Correct caller name (62c7df9)

v2022.1.0 (2022-12-30)

Feature

  • Add equal weight portfolio pipeline (a2a7db6)

  • Add bias statistics (efe3f46)

  • Add standardized returns in bias check (128c3f2)

  • Get covariance and correlation from risk model (02b606d)

  • Support transforming rolling factor risk model (77810b0)

  • Support exporting rolling factor risk model (5da7006)

  • Support factor risk model transformer (e26b61c)

  • Support factor risk model transformer (770ba6a)

  • Add factor covariance attribute in risk model (0b15a40)

  • Add WLS regressors (2b8d60b)

  • Change risk model input format to indexed on date / time rather than instruments (267d9a4)

  • Add parameter to show progress in rolling PCA (35fc9e3)

  • Add statistical risk model Rolling PCA (603344c)

  • Add the first statistical model PCA (4da36f7)

Fix

  • Incorrect mapping of factor exposures (4bc0028)

Documentation

[Unreleased]