Merge pull request #35 from atiqm/master

feat: Add TransferForestClassifier

Author: antoinedemathelin

.github/workflows/check-docs.yml

@@ -18,7 +18,7 @@ jobs:
       run: |
         sudo apt install pandoc
         python -m pip install --upgrade pip
-        pip install sphinx numpydoc nbsphinx sphinx_gallery sphinx_rtd_theme ipython
+        pip install jinja2==3.0.3 sphinx numpydoc nbsphinx sphinx_gallery sphinx_rtd_theme ipython
     - name: Install adapt dependencies
       run: |
         python -m pip install --upgrade pip

.github/workflows/publish-doc-to-remote.yml

@@ -18,7 +18,7 @@ jobs:
       run: |
         sudo apt install pandoc
         python -m pip install --upgrade pip
-        pip install sphinx numpydoc nbsphinx sphinx_gallery sphinx_rtd_theme ipython
+        pip install jinja2==3.0.3 sphinx numpydoc nbsphinx sphinx_gallery sphinx_rtd_theme ipython
     - name: Install adapt dependencies
       run: |
         python -m pip install --upgrade pip

adapt/_tree_utils.py

@@ -1,6 +1,31 @@
 import copy
 import numpy as np
 
+
+def _bootstrap_(size,class_wise=False,y=None):   
+    if class_wise:
+        if y is None:
+            print("Error : need labels to apply class wise bootstrap.")
+        else:
+            inds = []
+            oob_inds = []
+            classes_ = set(y)
+            ind_classes_ = np.zeros(len(classes_),dtype=object)
+            
+            for j,c in enumerate(classes_):
+                ind_classes_[j] = np.where(y==c)[0]
+                s = ind_classes_[j].size
+                inds += list(np.random.choice(ind_classes_[j], s, replace=True))
+                oob_inds += list(set(ind_classes_[j]) - set(inds))
+                
+            inds,oob_inds = np.array(inds),np.array(oob_inds)
+    else:
+        inds = np.random.choice(np.arange(size), size, replace=True)
+        oob_inds = set(np.arange(size)) - set(inds)
+        oob_inds = np.array(list(oob_inds))
+    
+    return inds, oob_inds
+    
 def depth_tree(dt,node=0):
 
     if dt.tree_.feature[node] == -2:
@@ -464,8 +489,7 @@ def coherent_new_split(phi,th,rule):
             return 0,1
         else:
             return 1,0
-
-
+      
 def all_coherent_splits(rule,all_splits):
 
     inds = np.zeros(all_splits.shape[0],dtype=bool)
@@ -534,4 +558,3 @@ def bounds_rule(rule,n_features):
 
     return bound_infs,bound_sups
 
-

adapt/parameter_based/__init__.py

@@ -5,9 +5,11 @@
 from ._regular import RegularTransferLR, RegularTransferLC, RegularTransferNN
 from ._finetuning import FineTuning
 from ._transfer_tree import TransferTreeClassifier
+from ._transfer_tree import TransferForestClassifier
 
 __all__ = ["RegularTransferLR",
            "RegularTransferLC",
            "RegularTransferNN",
            "FineTuning",
-           "TransferTreeClassifier"]
+           "TransferTreeClassifier",
+           "TransferForestClassifier"]