Introduction tutorial

hierarchicalforecast/core.py

@@ -18,9 +18,17 @@
 def _build_fn_name(fn) -> str:
     fn_name = type(fn).__name__
     func_params = fn.__dict__
+
+    # Take default parameter out of names
     args_to_remove = ['insample']
     if not func_params.get('nonnegative', False):
         args_to_remove += ['nonnegative']
+
+    if fn_name == 'MinTrace' and \
+        func_params['method']=='mint_shrink':
+        if func_params['mint_shr_ridge'] == 2e-8:
+            args_to_remove += ['mint_shr_ridge']
+
     func_params = [f'{name}-{value}' for name, value in func_params.items() if name not in args_to_remove]
     if func_params:
         fn_name += '_' + '_'.join(func_params)
@@ -58,12 +66,12 @@ def reconcile(self,
 
         The `reconcile` method is analogous to SKLearn `fit` method, it applies different 
         reconciliation methods instantiated in the `reconcilers` list. 
-        
+
         Most reconciliation methods can be described by the following convenient 
         linear algebra notation:
 
         $$\\tilde{\mathbf{y}}_{[a,b],\\tau} = \mathbf{S}_{[a,b][b]} \mathbf{P}_{[b][a,b]} \hat{\mathbf{y}}_{[a,b],\\tau}$$
-        
+
         where $a, b$ represent the aggregate and bottom levels, $\mathbf{S}_{[a,b][b]}$ contains
         the hierarchical aggregation constraints, and $\mathbf{P}_{[b][a,b]}$ varies across 
         reconciliation methods. The reconciled predictions are $\\tilde{\mathbf{y}}_{[a,b],\\tau}$, and the 
@@ -77,7 +85,7 @@ def reconcile(self,
         `tags`: Each key is a level and its value contains tags associated to that level.<br>
         `level`: float list 0-100, confidence levels for prediction intervals.<br>
         `intervals_method`: str, method used to calculate prediction intervals, one of `normality`, `bootstrap`, `permbu`.<br>
-         
+
         **Returns:**<br>
         `y_tilde`: pd.DataFrame, with reconciled predictions.        
         """

hierarchicalforecast/methods.py

@@ -368,7 +368,7 @@ class MinTrace:
     **Parameters:**<br>
     `method`: str, one of `ols`, `wls_struct`, `wls_var`, `mint_shrink`, `mint_cov`.<br>
     `nonnegative`: bool, reconciled forecasts should be nonnegative?<br>
-    `mint_shr_ridge`: float, ridge numeric protection to MinTrace-shr covariance estimator.<br>
+    `mint_shr_ridge`: float=2e-8, ridge numeric protection to MinTrace-shr covariance estimator.<br>
 
     **References:**<br>
     - [Wickramasuriya, S. L., Athanasopoulos, G., & Hyndman, R. J. (2019). \"Optimal forecast reconciliation for

nbs/core.ipynb

@@ -64,9 +64,17 @@
     "def _build_fn_name(fn) -> str:\n",
     "    fn_name = type(fn).__name__\n",
     "    func_params = fn.__dict__\n",
+    "\n",
+    "    # Take default parameter out of names\n",
     "    args_to_remove = ['insample']\n",
     "    if not func_params.get('nonnegative', False):\n",
     "        args_to_remove += ['nonnegative']\n",
+    "\n",
+    "    if fn_name == 'MinTrace' and \\\n",
+    "        func_params['method']=='mint_shrink':\n",
+    "        if func_params['mint_shr_ridge'] == 2e-8:\n",
+    "            args_to_remove += ['mint_shr_ridge']\n",
+    "\n",
     "    func_params = [f'{name}-{value}' for name, value in func_params.items() if name not in args_to_remove]\n",
     "    if func_params:\n",
     "        fn_name += '_' + '_'.join(func_params)\n",
@@ -99,6 +107,10 @@
     "test_eq(\n",
     "    _build_fn_name(MinTrace(method='ols', nonnegative=True)), \n",
     "    'MinTrace_method-ols_nonnegative-True'\n",
+    ")\n",
+    "test_eq(\n",
+    "    _build_fn_name(MinTrace(method='mint_shr')), \n",
+    "    'MinTrace_method-mint_shr'\n",
     ")"
    ]
   },
@@ -147,12 +159,12 @@
     "\n",
     "        The `reconcile` method is analogous to SKLearn `fit` method, it applies different \n",
     "        reconciliation methods instantiated in the `reconcilers` list. \n",
-    "        \n",
+    "\n",
     "        Most reconciliation methods can be described by the following convenient \n",
     "        linear algebra notation:\n",
     "\n",
     "        $$\\\\tilde{\\mathbf{y}}_{[a,b],\\\\tau} = \\mathbf{S}_{[a,b][b]} \\mathbf{P}_{[b][a,b]} \\hat{\\mathbf{y}}_{[a,b],\\\\tau}$$\n",
-    "        \n",
+    "\n",
     "        where $a, b$ represent the aggregate and bottom levels, $\\mathbf{S}_{[a,b][b]}$ contains\n",
     "        the hierarchical aggregation constraints, and $\\mathbf{P}_{[b][a,b]}$ varies across \n",
     "        reconciliation methods. The reconciled predictions are $\\\\tilde{\\mathbf{y}}_{[a,b],\\\\tau}$, and the \n",
@@ -166,7 +178,7 @@
     "        `tags`: Each key is a level and its value contains tags associated to that level.<br>\n",
     "        `level`: float list 0-100, confidence levels for prediction intervals.<br>\n",
     "        `intervals_method`: str, method used to calculate prediction intervals, one of `normality`, `bootstrap`, `permbu`.<br>\n",
-    "         \n",
+    "\n",
     "        **Returns:**<br>\n",
     "        `y_tilde`: pd.DataFrame, with reconciled predictions.        \n",
     "        \"\"\"\n",
@@ -758,16 +770,18 @@
     "fcst = StatsForecast(df=Y_train_df,\n",
     "                     #models=[ETS(season_length=12), Naive()],\n",
     "                     models=[Naive()],\n",
-    "                     freq='Q', n_jobs=-1) \n",
-    "Y_hat_df = fcst.forecast(h=4)\n",
+    "                     freq='Q', n_jobs=-1)\n",
+    "Y_hat_df = fcst.forecast(h=4, fitted=True)\n",
+    "Y_fitted_df = fcst.forecast_fitted_values()\n",
     "\n",
     "# Reconcile the base predictions\n",
     "Y_train_df = Y_train_df.reset_index().set_index('unique_id')\n",
     "Y_hat_df = Y_hat_df.reset_index().set_index('unique_id')\n",
     "reconcilers = [BottomUp(),\n",
-    "               MinTrace(method='ols')]\n",
+    "               MinTrace(method='mint_shrink')]\n",
     "hrec = HierarchicalReconciliation(reconcilers=reconcilers)\n",
-    "Y_rec_df = hrec.reconcile(Y_hat_df=Y_hat_df, Y_df=Y_train_df,\n",
+    "Y_rec_df = hrec.reconcile(Y_hat_df=Y_hat_df, \n",
+    "                          Y_df=Y_fitted_df,\n",
     "                          S=S_df, tags=tags)\n",
     "Y_rec_df.groupby('unique_id').head(2)"
    ]

nbs/examples/AustralianDomesticTourism.ipynb

@@ -975,7 +975,7 @@
     "\n",
     "Observe that we can recover the results reported by the [Forecasting: Principles and Practice](https://otexts.com/fpp3/tourism.html). The original results were calculated using the R package [fable](https://github.com/tidyverts/fable).\n",
     "\n",
-    "![image.png](./AustralianDomesticTourism-results-fable.png)"
+    "![Fable's reconciliation results](./imgs/AustralianDomesticTourism-results-fable.png)"
    ]
   },
   {

nbs/examples/AustralianPrisonPopulation.ipynb

@@ -769,7 +769,7 @@
     "\n",
     "Observe that we can recover the results reported by the [Forecasting: Principles and Practice](https://otexts.com/fpp3/prison.html) book. The original results were calculated using the R package [fable](https://github.com/tidyverts/fable).\n",
     "\n",
-    "![image](./AustralianPrisonPopulation-results-fable.png)"
+    "![Fable's reconciliation results](./imgs/AustralianPrisonPopulation-results-fable.png)"
    ]
   },
   {