Version 6.2.0

- Improvement in calculation speed of duplication_matrix, duplication_elimination_matrix and duplication_summation_matrix functions using a vectorized formula.
- Fixed formulation of risk parity with risk factors model that produced incorrect results when using the MOSEK solver.
- Fixed some bugs in PlotFunctions module.
- Fixed some bugs in HCPortfolio related to custom_mu vector and use of Kurtosis and Semi Kurtosis as risk measures.
- Standardized the way additional parameters to estimate mean vector and covariance matrix are entered.

Author: dcajasn

.github/workflows/build.yml

@@ -35,20 +35,8 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        os: [ ubuntu-20.04, macos-11, windows-2022 ]
-        python-version: ["3.8", "3.9", "3.10", "3.11", "3.12"]
-#        include:
-#          - os: ubuntu-20.04
-#            python-version: 3.8
-#            openmp: "True"
-#          - os: ubuntu-20.04
-#            python-version: 3.8
-#          - os: macos-11
-#            python-version: 3.8
-#            single_action_config: "True"
-#          - os: windows-2022
-#            python-version: 3.8
-
+        os: [ ubuntu-20.04, macos-12, windows-2022 ]
+        python-version: ["3.9", "3.10", "3.11", "3.12"]
 
     env:
       RUNNER_OS: ${{ matrix.os }}
@@ -90,16 +78,8 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        os: [ ubuntu-20.04, macos-11, windows-2022 ]
-        python-version: ["3.8", "3.9", "3.10", "3.11", "3.12"]
-#        include:
-#          - os: ubuntu-20.04
-#            python-version: 3.8
-#          - os: macos-11
-#            python-version: 3.8
-#            single_action_config: "True"
-#          - os: windows-2022
-#            python-version: 3.8
+        os: [ ubuntu-20.04, macos-12, windows-2022 ]
+        python-version: ["3.9", "3.10", "3.11", "3.12"]
 
     env:
       RUNNER_OS: ${{ matrix.os }}

CHANGELOG.rst

@@ -14,6 +14,14 @@ Changelog
     <a href='https://ko-fi.com/B0B833SXD' target='_blank'><img height='36'style='border:0px;height:36px;' src='https://cdn.ko-fi.com/cdn/kofi1.png?v=2' border='0' alt='Buy Me a Coffee at ko-fi.com' /></a>
 
 
+Version 6.2.0
+=============
+
+- Improvement in calculation speed of duplication_matrix, duplication_elimination_matrix and duplication_summation_matrix functions using a vectorized formula.
+- Fixed formulation of risk parity with risk factors model that produced incorrect results when using the MOSEK solver.
+- Fixed some bugs in PlotFunctions module.
+- Fixed some bugs in HCPortfolio related to custom_mu vector and use of Kurtosis and Semi Kurtosis as risk measures.
+- Standardized the way additional parameters to estimate mean vector and covariance matrix are entered.
 
 Version 6.1.0
 =============

README.md

@@ -163,9 +163,11 @@ Some of key functionalities that Riskfolio-Lib offers:
 - Tools to build views on assets and asset classes.
 - Tools to build views on risk factors.
 - Tools to build risk contribution constraints per asset classes.
+- Tools to build risk contribution constraints per risk factor using explicit risk factors and principal components.
 - Tools to build bounds constraints for Hierarchical Clustering Portfolios.
 - Tools to calculate risk measures.
 - Tools to calculate risk contributions per asset.
+- Tools to calculate risk contributions per risk factor.
 - Tools to calculate uncertainty sets for mean vector and covariance matrix.
 - Tools to calculate assets clusters based on codependence metrics.
 - Tools to estimate loadings matrix (Stepwise Regression and Principal Components Regression).
@@ -230,18 +232,18 @@ POW - refers to problems with 3-dimensional power cone constraints.
 
 ## Dependencies
 
-Riskfolio-Lib supports Python 3.8|.
+Riskfolio-Lib supports Python 3.9 or higher.
 
 Installation requires:
 - [numpy](http://www.numpy.org/) >= 1.24.0
 - [scipy](https://www.scipy.org/) >= 1.10.0
 - [pandas](https://pandas.pydata.org/) >= 2.0.0
-- [matplotlib](https://matplotlib.org/) >= 3.7.0
+- [matplotlib](https://matplotlib.org/) >= 3.8.0
 - [clarabel](https://oxfordcontrol.github.io/ClarabelDocs/stable/) >= 0.6.0
-- [cvxpy](https://www.cvxpy.org/) >= 1.4.0
+- [cvxpy](https://www.cvxpy.org/) >= 1.5.2
 - [scikit-learn](https://scikit-learn.org/stable/) >= 1.3.0
 - [statsmodels](https://www.statsmodels.org/) >= 0.13.5
-- [arch](https://bashtage.github.io/arch/) >= 5.4
+- [arch](https://bashtage.github.io/arch/) >= 7.0
 - [xlsxwriter](https://xlsxwriter.readthedocs.io) >= 3.1.2
 - [networkx](https://networkx.org) >= 3.0
 - [astropy](https://www.astropy.org) >= 5.1
@@ -260,7 +262,7 @@ If you use Riskfolio-Lib for published work, please use the following BibTeX ent
 ```
 @misc{riskfolio,
       author = {Dany Cajas},
-      title = {Riskfolio-Lib (6.1.0)},
+      title = {Riskfolio-Lib (6.2.0)},
       year  = {2024},
       url   = {https://github.com/dcajasn/Riskfolio-Lib},
       }

docs/source/conf.py

@@ -25,7 +25,7 @@
 copyright = '2020-2024, Dany Cajas'
 author = 'Dany Cajas'
 
-__version__ = "6.1.1"
+__version__ = "6.2.0"
 
 # The short X.Y version
 version = '.'.join(__version__.split('.')[:2])

docs/source/examples.rst

@@ -79,15 +79,17 @@ Hierarchical Clustering Portfolio Optimization
 * `Hierarchical Equal Risk Contribution (HERC) with Constraints <https://nbviewer.jupyter.org/github/dcajasn/Riskfolio-Lib/blob/master/examples/Tutorial%2044%20-%20Hierarchical%20Equal%20Risk%20Contribution%20(HERC)%20Portfolio%20Optimization%20with%20Constraints.ipynb>`_.
 * `Nested Clustered Optimization (NCO) with Constraints <https://nbviewer.jupyter.org/github/dcajasn/Riskfolio-Lib/blob/master/examples/Tutorial%2045%20-%20Nested%20Clustered%20Optimization%20(NCO)%20Portfolio%20Optimization%20with%20Constraints.ipynb>`_.
 
+
 Graph Theory Constraints
 ------------------------
 * `Hierarchical Clustering and Networks <https://nbviewer.jupyter.org/github/dcajasn/Riskfolio-Lib/blob/master/examples/Tutorial%2028%20-%20Hierarchical%20Clustering%20and%20Networks.ipynb>`_.
 * `Classic Mean Risk Optimization with Network and Dendrogram Constraints <https://nbviewer.jupyter.org/github/dcajasn/Riskfolio-Lib/blob/master/examples/Tutorial%2046%20-%20Classic%20Mean%20Risk%20Optimization%20with%20Network%20and%20Dendrogram%20Constraints.ipynb>`_.
 
+
 Backtesting
 -----------
-* `Multi Assets Algorithmic Trading Backtesting with transaction cost and slippage using Backtrader <https://nbviewer.jupyter.org/github/dcajasn/Riskfolio-Lib/blob/master/examples/Tutorial%205%20-%20Multi%20Assets%20Algorithmic%20Trading%20Backtesting%20with%20Backtrader.ipynb>`_ (matplotlib=3.2.2 for compatibility with backtrader=1.9.76.123).
-* `Multi Assets Algorithmic Trading Backtesting with Vectorbt <https://nbviewer.jupyter.org/github/dcajasn/Riskfolio-Lib/blob/master/examples/Tutorial%2018%20-%20Multi%20Assets%20Algorithmic%20Trading%20Backtesting%20with%20Vectorbt.ipynb>`_ (vectorbt=0.23.0).
+* `Multi Assets Algorithmic Trading Backtesting with transaction cost and slippage using Backtrader <https://nbviewer.jupyter.org/github/dcajasn/Riskfolio-Lib/blob/master/examples/Tutorial%205%20-%20Multi%20Assets%20Algorithmic%20Trading%20Backtesting%20with%20Backtrader.ipynb>`_ (matplotlib=3.2.2 for compatibility with backtrader=1.9.76.123. We don't recommend to try to reproduce this example due the compatibility problems of Backtrader).
+* `Multi Assets Algorithmic Trading Backtesting with Vectorbt <https://nbviewer.jupyter.org/github/dcajasn/Riskfolio-Lib/blob/master/examples/Tutorial%2018%20-%20Multi%20Assets%20Algorithmic%20Trading%20Backtesting%20with%20Vectorbt.ipynb>`_ (vectorbt=0.26.1).
 
 
 Excel and Reporting

docs/source/index.rst

@@ -44,9 +44,10 @@ Riskfolio-Lib
 Description
 ===========
 
-Riskfolio-Lib is a library for making portfolio optimization and quantitative strategic asset allocation in Python made in Peru |:peru:|. Its objective is to help students, academics and practitioners to build investment portfolios based on mathematically complex models with low effort. It is built on top of
-`CVXPY <https://www.cvxpy.org/>`_ and closely integrated
-with `Pandas <https://pandas.pydata.org/>`_ data structures.
+Riskfolio-Lib is a library for making portfolio optimization and quantitative strategic asset allocation 
+in Python made in Peru |:peru:|. Its objective is to help students, academics and practitioners to build 
+investment portfolios based on mathematically complex models with low effort. It is built on top of
+`CVXPY <https://www.cvxpy.org/>`_ and closely integrated with `Pandas <https://pandas.pydata.org/>`_ data structures.
 
 Some of key functionalities that Riskfolio-Lib offers:
 
@@ -178,9 +179,11 @@ Some of key functionalities that Riskfolio-Lib offers:
 - Tools to build views on assets and asset classes.
 - Tools to build views on risk factors.
 - Tools to build risk contribution constraints per asset classes.
+- Tools to build risk contribution constraints per risk factor using explicit risk factors and principal components.
 - Tools to build bounds constraints for Hierarchical Clustering Portfolios.
 - Tools to calculate risk measures.
 - Tools to calculate risk contributions per asset.
+- Tools to calculate risk contributions per risk factor.
 - Tools to calculate uncertainty sets for mean vector and covariance matrix.
 - Tools to calculate assets clusters based on codependence metrics.
 - Tools to estimate loadings matrix (Stepwise Regression and Principal Components Regression).
@@ -307,7 +310,7 @@ If you use Riskfolio-Lib for published work, please use the following BibTeX ent
 
     @misc{riskfolio,
           author = {Dany Cajas},
-          title = {Riskfolio-Lib (6.1.0)},
+          title = {Riskfolio-Lib (6.2.0)},
           year  = {2024},
           url   = {https://github.com/dcajasn/Riskfolio-Lib},
           }

docs/source/install.rst

@@ -69,15 +69,15 @@ Dependencies
 
 Riskfolio-Lib has the following dependencies:
 
-* numpy>=1.18.0
-* scipy>=1.1.0
-* pandas>=1.0.0
-* matplotlib>=3.5.0
-* cvxpy>=1.3.1
-* scikit-learn>=1.0.0
-* statsmodels>=0.10.1
-* arch>=4.15
-* xlsxwriter>=1.3.7
-* networkx>=2.5.1
-* astropy>=4.3.1 (if there are problems check `astropy installation instructions <https://www.astropy.org>`_)
+* numpy>=1.24.0
+* pandas>=2.0.0
+* matplotlib>=3.8.0
+* clarabel>=0.6.0
+* cvxpy>=1.5.2
+* scikit-learn>=1.3.0
+* statsmodels>=0.13.5
+* arch>=7.0
+* xlsxwriter>=3.1.2
+* networkx>=3.0
+* astropy>=5.1 (if there are problems check `astropy installation instructions <https://www.astropy.org>`_)
 * pybind11>=2.10.1

docs/source/portfolio.rst

@@ -27,17 +27,18 @@ one of the following 4 objective functions:
 .. math::
     \begin{aligned}
     &\underset{w}{\max} & & R (w)\\
-    &\text{s.t.} & & Aw \geq B\\
+    &\text{s.t.} & & Aw \leq B\\
     & & &\phi_{i}(w) \leq c_{i} \; \forall \; i \; \in \; [1,13] \\
     & & & R (w) \geq \overline{\mu}
     \end{aligned}
 
+
 - **Minimum Risk Portfolio:**
 
 .. math::
     \begin{aligned}
     &\underset{w}{\min} & & \phi_{k}(w)\\
-    &\text{s.t.} & & Aw \geq B\\
+    &\text{s.t.} & & Aw \leq B\\
     & & &\phi_{i}(w) \leq c_{i} \; \forall \; i \; \in \; [1,13] \\
     & & & R (w) \geq \overline{\mu}
     \end{aligned}
@@ -48,7 +49,7 @@ one of the following 4 objective functions:
 .. math::
     \begin{aligned}
     &\underset{w}{\max} & & \frac{R (w) - r_{f}}{\phi_{k}(w)}\\
-    &\text{s.t.} & & Aw \geq B\\
+    &\text{s.t.} & & Aw \leq B\\
     & & &\phi_{i}(w) \leq c_{i} \; \forall \; i \; \in \; [1,13] \\
     & & & R (w) \geq \overline{\mu}
     \end{aligned}
@@ -59,13 +60,13 @@ one of the following 4 objective functions:
 .. math::
     \begin{aligned}
     &\underset{w}{\max} & & R (w) - \lambda \phi_{k}(w)\\
-    &\text{s.t.} & & Aw \geq B\\
+    &\text{s.t.} & & Aw \leq B\\
     & & &\phi_{i}(w) \leq c_{i} \; \forall \; i \; \in \; [1,13] \\
     & & & R (w) \geq \overline{\mu}
     \end{aligned}
 
 
-Where:
+where:
 
 :math:`R (w)` is the return function, posible values are:
 
@@ -81,7 +82,7 @@ Where:
 
 :math:`r`: is the matrix of assets returns.
 
-:math:`Aw \geq B`: is a set of linear constraints.
+:math:`Aw \leq B`: is a set of linear constraints.
 
 :math:`\phi_{i}(w)`: are 20 available risk measures. The available risk
 measures are:
@@ -128,20 +129,20 @@ the general vanilla risk parity model :cite:`a-Roncalli` :cite:`a-RichardRoncall
     &\underset{w}{\min} & & \phi(w) \\
     &\text{s.t.} & & b \log(w) \geq c \\
     & & & \mu w \geq \overline{\mu} \\
-    & & & Aw \geq B \\
+    & & & Aw \leq B \\
     & & & w \geq 0 \\
     \end{aligned}
 
 
-Where:
+where:
 
 :math:`w`: is the vector of weights of the optimum portfolio.
 
 :math:`\mu`: is the vector of expected returns.
 
 :math:`b`: is a vector of risk contribution constraints.
 
-:math:`Aw \geq B`: is a set of linear constraints.
+:math:`Aw \leq B`: is a set of linear constraints.
 
 :math:`\phi(w)`: are 16 available risk measures. The available risk
 measures are:
@@ -182,13 +183,13 @@ the relaxed risk parity model :cite:`a-GambetaKwon`:
     & & & w_{i} \zeta_{i} \geq \gamma^{2} b_{i} & \forall i=1 , \ldots , N \\
     & & & \lambda x^{T} \Theta x \leq \rho^{2} & \\
     & & & \mu w \geq \overline{\mu} & \\
-    & & & Aw \geq B & \\
+    & & & Aw \leq B & \\
     & & & \sum^{N}_{i=1} w_{i} = 1 & \\
     & & & \psi, \gamma, \rho, w  \geq 0 & \\
     \end{aligned}
 
 
-Where:
+where:
 
 :math:`w`: is the vector of weights of the optimum portfolio.
 
@@ -210,7 +211,7 @@ Where:
 
 :math:`\Theta = \text{diag}(\Sigma)`
 
-:math:`Aw \geq B`: is a set of linear constraints.
+:math:`Aw \leq B`: is a set of linear constraints.
 
 
 Worst Case Mean Variance Portfolio Optimization
@@ -226,7 +227,7 @@ that results from optimize one of the following 4 objective functions:
 .. math::
     \begin{aligned}
     &\underset{w}{\max} & & \underset{\mu \, \in \, U_{\mu}}{\min} \mu w\\
-    &\text{s.t.} & & Aw \geq B\\
+    &\text{s.t.} & & Aw \leq B\\
     \end{aligned}
 
 
@@ -235,7 +236,7 @@ that results from optimize one of the following 4 objective functions:
 .. math::
     \begin{aligned}
     &\underset{w}{\min} & & \underset{\Sigma \, \in \, U_{\Sigma}}{\max} w^{T} \Sigma w\\
-    &\text{s.t.} & & Aw \geq B\\
+    &\text{s.t.} & & Aw \leq B\\
     \end{aligned}
 
 
@@ -245,7 +246,7 @@ that results from optimize one of the following 4 objective functions:
     \begin{aligned}
     &\underset{w}{\max} & & \cfrac{\underset{\mu \, \in \, U_{\mu}}{\min} \mu w - r_{f}}
     {\underset{\Sigma \, \in \, U_{\Sigma}}{\max} \sqrt{w^{T} \Sigma w}}\\
-    &\text{s.t.} & & Aw \geq B\\
+    &\text{s.t.} & & Aw \leq B\\
     \end{aligned}
 
 
@@ -255,11 +256,11 @@ that results from optimize one of the following 4 objective functions:
     \begin{aligned}
     &\underset{w}{\max} & & \underset{\mu \, \in \, U_{\mu}}{\min} \mu w
     - \underset{\Sigma \, \in \, U_{\Sigma}}{\max}  \lambda w^{T} \Sigma w\\
-    &\text{s.t.} & & Aw \geq B\\
+    &\text{s.t.} & & Aw \leq B\\
     \end{aligned}
 
 
-Where:
+where:
 
 :math:`w` are the weights of the portfolio.
 
@@ -285,7 +286,7 @@ Where:
     \end{aligned}
 
 
-:math:`Aw \geq B`: is a set of linear constraints.
+:math:`Aw \leq B`: is a set of linear constraints.
 
 :math:`r_{f}`: is the risk free rate. 
 
@@ -303,7 +304,7 @@ this function to calculate the Higher L-Moment portfolio optimization model :cit
 .. math::
     \begin{aligned}
     &\underset{w}{\min} & & \sum^{T}_{i=0} v_{[i]}y_{[i]} \\
-    &\text{s.t.} & & Aw \geq B\\
+    &\text{s.t.} & & Aw \leq B\\
     & & & y = rw \\
     & & & R (w) \geq \overline{\mu} \\
     \end{aligned}
@@ -313,7 +314,7 @@ this function to calculate the Higher L-Moment portfolio optimization model :cit
 .. math::
     \begin{aligned}
     &\underset{w}{\max} & & \frac{R (w) - r_{f}}{\sum^{T}_{i=0} v_{[i]}y_{[i]}}\\
-    &\text{s.t.} & & Aw \geq B\\
+    &\text{s.t.} & & Aw \leq B\\
     & & & y = rw \\
     & & & R (w) \geq \overline{\mu} \\
     \end{aligned}
@@ -323,7 +324,7 @@ this function to calculate the Higher L-Moment portfolio optimization model :cit
 .. math::
     \begin{aligned}
     &\underset{w}{\max} & & R (w) - \lambda \left ( \sum^{T}_{i=0} v_{[i]}y_{[i]} \right) \\
-    &\text{s.t.} & & Aw \geq B\\
+    &\text{s.t.} & & Aw \leq B\\
     & & & y = rw \\
     & & & R (w) \geq \overline{\mu} \\
     \end{aligned}