Merge pull request #20 from skfolio/fix/issue-11/business-days-in-year

[BUG] Default trading days in a year amended from 255 to 252

README.rst

@@ -388,7 +388,7 @@ Randomized Search of the L2 Norm
 
     randomized_search = RandomizedSearchCV(
         estimator=MeanRisk(),
-        cv=WalkForward(train_size=255, test_size=60),
+        cv=WalkForward(train_size=252, test_size=60),
         param_distributions={
             "l2_coef": loguniform(1e-3, 1e-1),
         },

docs/conf.py

@@ -26,6 +26,8 @@
     ),
 )
 warnings.filterwarnings("ignore", category=UserWarning)
+warnings.filterwarnings("ignore", category=FutureWarning)
+
 # -- Project information -----------------------------------------------------
 
 project = "skfolio"

docs/index.rst

@@ -332,7 +332,7 @@ Randomized Search of the L2 Norm
 
     randomized_search = RandomizedSearchCV(
         estimator=MeanRisk(),
-        cv=WalkForward(train_size=255, test_size=60),
+        cv=WalkForward(train_size=252, test_size=60),
         param_distributions={
             "l2_coef": loguniform(1e-3, 1e-1),
         },

docs/user_guide/data_preparation.rst

@@ -52,7 +52,7 @@ It is not uncommon to witness the following steps [1]_:
 #. Take the daily prices :math:`S_{t}, S_{t+1}, ...,` for all the n securities
 #. Transform the daily prices to daily logarithmic returns
 #. Estimate the expected returns vector :math:`\mu` and covariance matrix :math:`\Sigma` from the daily logarithmic returns
-#. Determine the investment horizon, for example k = 255 days
+#. Determine the investment horizon, for example k = 252 days
 #. Project the expected returns and covariance to the horizon using the square-root rule: :math:`\mu_{k} ≡ k \times \mu` and :math:`\Sigma_{k} ≡ k \times \Sigma`
 #. Compute the mean-variance efficient frontier :math:`\max_{w} \Biggl\{ w^T \mu - \lambda \times w^T \Sigma w \Biggr\}`
 
@@ -82,10 +82,10 @@ Example for stocks
 #. Take the prices :math:`S_{t}, S_{t+1}, ...,` (for example daily) for all the n securities
 #. Transform the daily prices to daily logarithmic returns. Note that linear return is also a market invariant for stock, however logarithmic return is going to simplify step 3) and 4).
 #. Estimate the joint distribution of market invariants by fitting parametrically the daily logarithmic returns to a multivariate normal distribution: estimate the joint distribution parameters :math:`\mu^{Log}_{daily}` and :math:`\Sigma^{Log}_{daily}`
-#. Project the distribution of invariants to the time period of investment (for example one year i.e. 255 business days). Because logarithmic returns are additive across time, we have:
+#. Project the distribution of invariants to the time period of investment (for example one year i.e. 252 business days). Because logarithmic returns are additive across time, we have:
 
-        * .. math:: \mu^{Log}_{yearly} = 255 \times \mu^{Log}_{daily}
-        * .. math:: \Sigma^{Log}_{yearly} = 255 \times \Sigma^{Log}_{daily}
+        * .. math:: \mu^{Log}_{yearly} = 252 \times \mu^{Log}_{daily}
+        * .. math:: \Sigma^{Log}_{yearly} = 252 \times \Sigma^{Log}_{daily}
 #. Compute the distribution of linear returns at the investment horizon. Using the characteristic function of the normal distribution, and the pricing function :math:`S_{yearly} = S_{0} e^{R^{Log}_{yearly}}`, we get:
 
         * .. math:: \mathbb{E}(S_{yearly}) = \pmb{s}_{0} \circ exp\Biggl(\pmb{\mu}^{Log}_{yearly} + \frac{1}{2} diag\Biggl(\pmb{\Sigma}^{Log}_{yearly}\Biggr)\Biggr)

examples/1_mean_risk/plot_3_efficient_frontier.py

@@ -89,7 +89,7 @@
 # 15%, 20%, 25%, 30% and 35% (annualized):
 model = MeanRisk(
     risk_measure=RiskMeasure.VARIANCE,
-    min_return=np.array([0.15, 0.20, 0.25, 0.30, 0.35]) / 255,
+    min_return=np.array([0.15, 0.20, 0.25, 0.30, 0.35]) / 252,
     portfolio_params=dict(name="Variance"),
 )
 

examples/1_mean_risk/plot_6_transaction_costs.py

@@ -51,7 +51,7 @@
     * Duration = 1 months (21 business days):
         * 1 month expected return A ≈ -0.8%
         * 1 month expected return B ≈ 0.1%
-    * Duration = 1 year (255 business days):
+    * Duration = 1 year (252 business days):
         * 1 year expected return A ≈ 1.5%
         * 1 year expected return B ≈ 1.3%
 

examples/1_mean_risk/plot_7_management_fees.py

@@ -75,9 +75,9 @@
 #   * JPMorgan: 1% p.a.
 #
 # The MF are expressed in per annum, so we need to convert them in daily MF.
-# We suppose 255 trading days in a year:
-management_fees = {"AAPL": 0.03 / 255, "GE": 0.06 / 255, "JPM": 0.01 / 255}
-# Same as management_fees = np.array([0.03, 0.06, 0.01]) / 255
+# We suppose 252 trading days in a year:
+management_fees = {"AAPL": 0.03 / 252, "GE": 0.06 / 252, "JPM": 0.01 / 252}
+# Same as management_fees = np.array([0.03, 0.06, 0.01]) / 252
 
 model_mf = MeanRisk(
     objective_function=ObjectiveFunction.MAXIMIZE_UTILITY,
@@ -101,7 +101,7 @@
 
 # %%
 # As explained above, we transform the yearly MF into a daily MF:
-management_fees = np.array([0.03, 0.06, 0.01]) / 255
+management_fees = np.array([0.03, 0.06, 0.01]) / 252
 
 # %%
 # First, we train the model without MF and test it with MF.

examples/1_mean_risk/plot_8_regularization.py

@@ -66,7 +66,7 @@
 model = MeanRisk(
     risk_measure=RiskMeasure.VARIANCE,
     min_weights=-1,
-    max_variance=0.3**2 / 255,
+    max_variance=0.3**2 / 252,
     efficient_frontier_size=30,
     portfolio_params=dict(name="Mean-Variance", tag="No Regularization"),
 )
@@ -78,7 +78,7 @@
 model_l1 = MeanRisk(
     risk_measure=RiskMeasure.VARIANCE,
     min_weights=-1,
-    max_variance=0.3**2 / 255,
+    max_variance=0.3**2 / 252,
     efficient_frontier_size=30,
     l1_coef=0.001,
     portfolio_params=dict(name="Mean-Variance", tag="L1 Regularization"),
@@ -153,7 +153,7 @@
 # Hyper-parameter Tuning
 # ======================
 # In this section, we consider a 3 months rolling (60 business days) long-short
-# allocation fitted on the preceding year of data (255 business days) that maximizes the
+# allocation fitted on the preceding year of data (252 business days) that maximizes the
 # return under a volatility constraint of 30% p.a.
 #
 # We use `GridSearchCV` to select the optimal L1 and L2 regularization coefficients on
@@ -165,11 +165,11 @@
 ref_model = MeanRisk(
     risk_measure=RiskMeasure.VARIANCE,
     objective_function=ObjectiveFunction.MAXIMIZE_RETURN,
-    max_variance=0.3**2 / 255,
+    max_variance=0.3**2 / 252,
     min_weights=-1,
 )
 
-cv = WalkForward(train_size=255, test_size=60)
+cv = WalkForward(train_size=252, test_size=60)
 
 grid_search = GridSearchCV(
     estimator=ref_model,
@@ -296,14 +296,14 @@
 # |
 #
 # The highest mean out-of-sample Sharpe Ratio is 1.55 and is achieved for a L2 coef of
-# 0.031.
+# 0.023.
 # Also note that without regularization, the mean train Sharpe Ratio is around
 # six time higher than the mean test Sharpe Ratio. That would be a clear indiction of
 # overfitting.
 #
 # Now, we analyze all three models on the test set. By using `cross_val_predict` with
 # `WalkForward`, we are able to compute efficiently the `MultiPeriodPortfolio`
-# composed of 60 days rolling portfolios fitted on the preceding 255 days:
+# composed of 60 days rolling portfolios fitted on the preceding 252 days:
 
 benchmark = EqualWeighted()
 pred_bench = cross_val_predict(benchmark, X_test, cv=cv)
@@ -320,8 +320,8 @@
 
 # %%
 # From the plot and the below summary, we can see that the un-regularized model is
-# overfitted and perform poorly on the test set. Its annualized volatility is 53%, which
-# is significantly above the model upper-bound of 30% and its Sharpe Ratio is 0.29 which
+# overfitted and perform poorly on the test set. Its annualized volatility is 54%, which
+# is significantly above the model upper-bound of 30% and its Sharpe Ratio is 0.32 which
 # is the lowest of all models.
 
 population.summary()

examples/1_mean_risk/plot_9_uncertainty_set.py

@@ -96,7 +96,7 @@
 # Hyper-Parameter Tuning
 # ======================
 # In this section, we consider a 3 months rolling (60 business days) long-short
-# allocation fitted on the preceding year of data (255 business days) that maximizes the
+# allocation fitted on the preceding year of data (252 business days) that maximizes the
 # portfolio return under a CVaR constraint.
 # We will use `GridSearchCV` to select the below model parameters on the training set
 # using walk forward analysis with a Mean/CVaR ratio scoring.
@@ -113,22 +113,23 @@
 model_no_uncertainty = MeanRisk(
     risk_measure=RiskMeasure.CVAR,
     objective_function=ObjectiveFunction.MAXIMIZE_RETURN,
-    max_cvar=0.04,
+    max_cvar=0.02,
+    cvar_beta=0.9,
     min_weights=-1,
 )
 
 model_uncertainty = clone(model_no_uncertainty)
 model_uncertainty.set_params(mu_uncertainty_set_estimator=EmpiricalMuUncertaintySet())
 
-cv = WalkForward(train_size=255, test_size=60)
+cv = WalkForward(train_size=252, test_size=60)
 
 grid_search = GridSearchCV(
     estimator=model_uncertainty,
     cv=cv,
     n_jobs=-1,
     param_grid={
         "mu_uncertainty_set_estimator__confidence_level": [0.80, 0.90],
-        "max_cvar": [0.04, 0.05, 0.06],
+        "max_cvar": [0.03, 0.04, 0.05],
         "cvar_beta": [0.8, 0.9, 0.95],
     },
     scoring=make_scorer(RatioMeasure.CVAR_RATIO),
@@ -138,7 +139,7 @@
 print(best_model)
 
 # %%
-# The optimal parameters among the above 2x3x3 grid are the `max_cvar=6%`,
+# The optimal parameters among the above 2x3x3 grid are the `max_cvar=3%`,
 # `cvar_beta=90%` and :class:`~skfolio.uncertainty_set.EmpiricalMuUncertaintySet`
 # `confidence_level=80%`. These parameters are the ones that achieved the highest mean
 # out-of-sample Mean/CVaR ratio.
@@ -204,7 +205,7 @@
 # Now, we analyze all three models on the test set.
 # By using `cross_val_predict` with `WalkForward`, we are able to compute efficiently
 # the `MultiPeriodPortfolio` composed of 60 days rolling portfolios fitted on the
-# preceding 255 days:
+# preceding 252 days:
 pred_no_uncertainty = cross_val_predict(model_no_uncertainty, X_test, cv=cv)
 pred_no_uncertainty.name = "No Uncertainty set"
 
@@ -219,7 +220,7 @@
 
 # %%
 # From the plot and the below summary, we can see that the model without uncertainty set
-# is overfitted and perform poorly on the test set. Its CVaR at 95% is 18% and its
+# is overfitted and perform poorly on the test set. Its CVaR at 95% is 10% and its
 # Mean/CVaR ratio is 0.006 which is the lowest of all models.
 population.summary()
 

examples/5_clustering/plot_1_hrp_cvar.py

@@ -100,7 +100,7 @@
 # The original HRP is based on the single-linkage (equivalent to the minimum spanning
 # tree), which suffers from the chaining effect.
 # In the :class:`~skfolio.optimization.HierarchicalRiskParity` estimator, the default
-# linkage method is set to the Ward variance minimization algorithm, which is more 
+# linkage method is set to the Ward variance minimization algorithm, which is more
 # stable and has better properties than the single-linkage method.
 #
 # However, since the HRP optimization doesn’t utilize the full cluster structure but

examples/5_clustering/plot_3_hrp_vs_herc.py

@@ -8,7 +8,7 @@
 :class:`~skfolio.optimization.HierarchicalEqualRiskContribution` (HERC) optimization.
 
 For that comparison, we consider a 3 months rolling (60 business days) allocation fitted
-on the preceding year of data (255 business days) that minimizes the CVaR.
+on the preceding year of data (252 business days) that minimizes the CVaR.
 
 We will employ `GridSearchCV` to select the optimal parameters of each model on the
 training set using cross-validation that achieves the highest average out-of-sample
@@ -71,8 +71,8 @@
 # For both HRP and HERC models, we find the parameters that maximizes the average
 # out-of-sample Mean-CVaR ratio using `GridSearchCV` with `WalkForward` cross-validation
 # on the training set. The `WalkForward` are chosen to simulate a three months
-# (60 business days) rolling portfolio fitted on the previous year (255 business days):
-cv = WalkForward(train_size=255, test_size=60)
+# (60 business days) rolling portfolio fitted on the previous year (252 business days):
+cv = WalkForward(train_size=252, test_size=60)
 
 grid_search_hrp = GridSearchCV(
     estimator=model_hrp,
@@ -155,7 +155,7 @@
 
 # %%
 # We choose `n_folds` and `n_test_folds` to obtain more than 100 test paths and an average
-# training size of approximately 255 days:
+# training size of approximately 252 days:
 cv.summary(X_test)
 
 # %%

examples/5_clustering/plot_4_nco.py

@@ -98,7 +98,7 @@
 model1.clustering_estimator_.plot_dendrogram()
 
 # %%
-# Linkage M
+# Linkage Methods
 # ===============
 # The hierarchical clustering can be greatly affected by the choice of the linkage
 # method. In the :class:`~skfolio.cluster.HierarchicalClustering` estimator, the default

examples/5_clustering/plot_5_nco_grid_search.py

@@ -57,8 +57,8 @@
 # We find the model parameters that maximizes the out-of-sample Sharpe ratio using
 # `GridSearchCV` with `WalkForward` cross-validation on the training set.
 # The `WalkForward` are chosen to simulate a three months (60 business days) rolling
-# portfolio fitted on the previous year (255 business days):
-cv = WalkForward(train_size=255, test_size=60)
+# portfolio fitted on the previous year (252 business days):
+cv = WalkForward(train_size=252, test_size=60)
 
 grid_search_hrp = GridSearchCV(
     estimator=model_nco,
@@ -137,7 +137,7 @@
 
 # %%
 # We choose `n_folds` and `n_test_folds` to obtain more than 30 test paths and an average
-# training size of approximately 255 days:
+# training size of approximately 252 days:
 cv.summary(X_test)
 
 # %%
@@ -157,9 +157,7 @@
 # Distribution
 # ============
 # We plot the out-of-sample distribution of Sharpe Ratio for the NCO model:
-pred_nco.plot_distribution(
-    measure_list=[RatioMeasure.ANNUALIZED_SHARPE_RATIO]
-)
+pred_nco.plot_distribution(measure_list=[RatioMeasure.ANNUALIZED_SHARPE_RATIO])
 
 # %%
 # Let's print the average and standard-deviation of out-of-sample Sharpe Ratios:

examples/7_pre_selection/plot_1_drop_correlated.py

@@ -99,15 +99,15 @@
     X_test,
     cv=cv,
     n_jobs=-1,
-    portfolio_params=dict(annualized_factor=255, tag="model1"),
+    portfolio_params=dict(annualized_factor=252, tag="model1"),
 )
 
 pred_2 = cross_val_predict(
     model2,
     X_test,
     cv=cv,
     n_jobs=-1,
-    portfolio_params=dict(annualized_factor=255, tag="model2"),
+    portfolio_params=dict(annualized_factor=252, tag="model2"),
 )
 
 # %%

examples/7_pre_selection/plot_2_select_best_performers.py

@@ -73,8 +73,8 @@
 # pre-selected assets `k` that maximizes the out-of-sample Sharpe Ratio using
 # `GridSearchCV` with `WalkForward` cross-validation on the training set. The
 # `WalkForward` is chosen to simulate a three months (60 business days) rolling
-# portfolio fitted on the previous year (255 business days):
-cv = WalkForward(train_size=255, test_size=60)
+# portfolio fitted on the previous year (252 business days):
+cv = WalkForward(train_size=252, test_size=60)
 
 scorer = make_scorer(RatioMeasure.ANNUALIZED_SHARPE_RATIO)
 # %%

examples/8_data_preparation/plot_1_investment_horizon.py

@@ -44,9 +44,9 @@
 population.extend(model.fit_predict(X))
 
 for tag, investment_horizon in [
-    ("3M", 255 / 4),
-    ("1Y", 255),
-    ("10Y", 10 * 255),
+    ("3M", 252 / 4),
+    ("1Y", 252),
+    ("10Y", 10 * 252),
 ]:
     model = MeanRisk(
         risk_measure=RiskMeasure.VARIANCE,

src/skfolio/cluster/_hierarchical.py

@@ -8,13 +8,13 @@
 
 import numpy as np
 import numpy.typing as npt
-import plotly.figure_factory as ff
 import plotly.graph_objects as go
 import scipy.cluster.hierarchy as sch
 import scipy.spatial.distance as scd
 import sklearn.base as skb
 import sklearn.utils.validation as skv
 
+from skfolio.utils.fixes import create_dendrogram
 from skfolio.utils.stats import assert_is_distance, compute_optimal_n_clusters
 from skfolio.utils.tools import AutoEnum, default_asset_names
 
@@ -245,7 +245,7 @@ def plot_dendrogram(self, heatmap: bool = True) -> go.Figure:
             asset_names = default_asset_names(n_assets=n_assets)
 
         if not heatmap:
-            fig = ff.create_dendrogram(
+            fig = create_dendrogram(
                 np.ones(1),
                 distfun=lambda x: None,
                 linkagefun=lambda x: linkage_matrix,
@@ -264,7 +264,7 @@ def plot_dendrogram(self, heatmap: bool = True) -> go.Figure:
             return fig
 
         # Initialize figure by creating upper dendrogram
-        fig = ff.create_dendrogram(
+        fig = create_dendrogram(
             np.ones(1),
             orientation="bottom",
             distfun=lambda x: None,
@@ -277,7 +277,7 @@ def plot_dendrogram(self, heatmap: bool = True) -> go.Figure:
             fig["data"][i]["yaxis"] = "y2"
 
         # Create Side Dendrogram
-        side_dendrogram = ff.create_dendrogram(
+        side_dendrogram = create_dendrogram(
             np.ones(1),
             orientation="right",
             distfun=lambda x: None,