Merge pull request #37 from Techtonique/polyak

Polyak

.gitignore

@@ -1,4 +1,7 @@
 mlsauce/.DS_Store
+*.html
+*.so 
+*.c
 
 # Byte-compiled / optimized / DLL files
 __pycache__/

Makefile

@@ -21,10 +21,10 @@ for line in sys.stdin:
 endef
 export PRINT_HELP_PYSCRIPT
 
-BROWSER := python -c "$$BROWSER_PYSCRIPT"
+BROWSER := python3 -c "$$BROWSER_PYSCRIPT"
 
 help:
-	@python -c "$$PRINT_HELP_PYSCRIPT" < $(MAKEFILE_LIST)
+	@python3 -c "$$PRINT_HELP_PYSCRIPT" < $(MAKEFILE_LIST)
 
 clean: clean-build clean-pyc clean-test ## remove all build, test, coverage and Python artifacts
 
@@ -34,7 +34,8 @@ clean-build: ## remove build artifacts
 	find . -name '*.egg-info' -exec rm -fr {} +
 	find . -name '*.egg' -exec rm -rf {} +
 	find . -name '*.so' -exec rm -f {} +
-	find . -name '*.c' -exec rm -f {} +	
+	find . -name '*.c' -exec rm -f {} +
+	find . -name '*.html' -exec rm -f {} +	
 
 clean-pyc: ## remove Python file artifacts
 	find . -name '*.pyc' -exec rm -f {} +
@@ -100,4 +101,4 @@ run-booster: ## run all boosting estimators examples with one command
 	find examples -maxdepth 2 -name "*boost_*.py" -exec  python3 {} \;
 
 run-lazy: ## run all lazy estimators examples with one command
-	find examples -maxdepth 2 -name "*lazy*.py" -exec  python3 {} \;
+	find examples -maxdepth 2 -name "*lazy*.py" -exec  python3 {} \;

examples/genboost_classifier.py

@@ -50,6 +50,8 @@
 
 print(obj.obj['loss'])
 
+print(obj.obj['fit_obj_i'])
+
 obj = ms.GenericBoostingClassifier(clf, tolerance=1e-2, n_clusters=2)
 print(obj.get_params())
 start = time()
@@ -61,6 +63,7 @@
 
 print(obj.obj['loss'])
 
+print(obj.obj['fit_obj_i'])
 
 # data 2
 print("\n")
@@ -84,6 +87,8 @@
 
 print(obj.obj['loss'])
 
+print(obj.obj['fit_obj_i'])
+
 obj = ms.GenericBoostingClassifier(clf, n_clusters=3)
 print(obj.get_params())
 start = time()
@@ -95,6 +100,8 @@
 
 print(obj.obj['loss'])
 
+print(obj.obj['fit_obj_i'])
+
 # data 3
 print("\n")
 print("iris data -----")
@@ -118,6 +125,8 @@
 
 print(obj.obj['loss'])
 
+print(obj.obj['fit_obj_i'])
+
 
 print("\n")
 print("GenericBoosting  KRR -----")
@@ -134,6 +143,8 @@
 
 print(obj.obj['loss'])
 
+print(obj.obj['fit_obj_i'])
+
 obj = ms.GenericBoostingClassifier(clf2, tolerance=1e-2, n_clusters=2)
 print(obj.get_params())
 start = time()
@@ -145,6 +156,8 @@
 
 print(obj.obj['loss'])
 
+print(obj.obj['fit_obj_i'])
+
 
 # data 2
 print("\n")
@@ -168,6 +181,8 @@
 
 print(obj.obj['loss'])
 
+print(obj.obj['fit_obj_i'])
+
 obj = ms.GenericBoostingClassifier(clf2, n_clusters=3)
 print(obj.get_params())
 start = time()
@@ -202,3 +217,5 @@
 
 print(obj.obj['loss'])
 
+print(obj.obj['fit_obj_i'])
+

examples/genboost_classifier_update.py

@@ -0,0 +1,86 @@
+import numpy as np 
+from sklearn.datasets import load_digits, load_breast_cancer, load_wine, load_iris
+from sklearn.model_selection import train_test_split, GridSearchCV, cross_val_score
+from sklearn.linear_model import Ridge, LinearRegression, RidgeCV, ElasticNetCV, LassoCV, Lasso
+from time import time
+from os import chdir
+from sklearn import metrics
+import os 
+
+print(f"\n ----- Running: {os.path.basename(__file__)}... ----- \n")
+
+print(os.path.relpath(os.path.dirname(__file__)))
+
+#wd="/workspace/mlsauce/mlsauce/examples"
+#
+#chdir(wd)
+
+import mlsauce as ms
+
+eta = 0.9
+
+
+# data 1
+regr1 = Ridge()
+
+
+print("\n Example 1 --------------------------- \n")
+wine = load_wine()
+X = wine.data
+y = wine.target
+# split data into training test and test set
+np.random.seed(15029)
+X_train, X_test, y_train, y_test = train_test_split(X, y, 
+                                                    test_size=0.2)
+
+obj = ms.GenericBoostingClassifier(regr1)
+print(obj.get_params())
+start = time()
+obj.fit(X_train, y_train)
+print(time()-start)
+print(f"obj.obj['fit_obj_i'][0]: {obj.obj['fit_obj_i'][0].coef_}")
+start = time()
+print(f"score init. {obj.score(X_test[3:,], 
+                y_test[3:])}")
+print(time()-start)
+obj = obj.update(X_test[0,:], y_test[0], eta=eta)
+obj = obj.update(X_test[1,:], y_test[1], eta=eta)
+obj = obj.update(X_test[2,:], y_test[2], eta=eta)
+print(f"score updated: {obj.score(X_test[3:,], 
+                y_test[3:])}")
+print(f"obj.obj['fit_obj_i'][0].coef_: {obj.obj['fit_obj_i'][0].coef_}")
+
+print("\n Example 2 --------------------------- \n")
+
+eta = 0.3
+
+regr2 = Ridge()
+
+dataset = load_iris()
+X = dataset.data
+y = dataset.target
+# split data into training test and test set
+X_train, X_test, y_train, y_test = train_test_split(X, y, 
+                                                    test_size=0.2, 
+                                                    random_state=13)
+
+obj = ms.GenericBoostingClassifier(regr2)
+print(obj.get_params())
+start = time()
+obj.fit(X_train, y_train)
+print(time()-start)
+print(f"obj.obj['fit_obj_i'][0]: {obj.obj['fit_obj_i'][0].coef_}")
+print(f"obj.obj['Ym']: {obj.obj['Ym']}")
+start = time()
+print(f"score init. {obj.score(X_test[5:,], 
+                y_test[5:])}")
+print(time()-start)
+obj = obj.update(X_test[0,:], y_test[0], eta=eta)
+obj = obj.update(X_test[1,:], y_test[1], eta=eta)
+obj = obj.update(X_test[2,:], y_test[2], eta=eta)
+obj = obj.update(X_test[3,:], y_test[3], eta=eta)
+obj = obj.update(X_test[4,:], y_test[4], eta=eta)
+print(f"score updated: {obj.score(X_test[5:,], 
+                y_test[5:])}")
+print(f"obj.obj['fit_obj_i'][0].coef_: {obj.obj['fit_obj_i'][0].coef_}")
+print(f"obj.obj['Ym']: {obj.obj['Ym']}")

examples/genboost_regressor.py

@@ -10,7 +10,8 @@
 import mlsauce as ms
 import numpy as np 
 import matplotlib.pyplot as plt
-from sklearn.datasets import load_diabetes
+from sklearn.datasets import load_diabetes, fetch_california_housing
+from sklearn.linear_model import Ridge, LinearRegression
 from sklearn.model_selection import train_test_split, GridSearchCV, cross_val_score
 from sklearn.tree import DecisionTreeRegressor
 from time import time
@@ -21,7 +22,7 @@
 print("\n")
 print("diabetes data -----")
 
-regr = DecisionTreeRegressor()
+regr = LinearRegression()
 
 diabetes = load_diabetes()
 X = diabetes.data
@@ -31,6 +32,8 @@
 X_train, X_test, y_train, y_test = train_test_split(X, y, 
                                                     test_size=0.2)
 
+print("\n Example 1 --------------------------- \n")
+
 obj = ms.GenericBoostingRegressor(regr, col_sample=0.9, row_sample=0.9)
 print(obj.get_params())
 start = time()
@@ -39,9 +42,10 @@
 start = time()
 print(np.sqrt(np.mean(np.square(obj.predict(X_test) - y_test))))
 print(time()-start)
-
 print(obj.obj['loss'])
 
+print("\n Example 2 --------------------------- \n")
+
 obj = ms.GenericBoostingRegressor(regr, col_sample=0.9, row_sample=0.9, n_clusters=2)
 print(obj.get_params())
 start = time()
@@ -50,5 +54,29 @@
 start = time()
 print(np.sqrt(np.mean(np.square(obj.predict(X_test) - y_test))))
 print(time()-start)
-
 print(obj.obj['loss'])
+print(obj.obj['fit_obj_i'])
+
+print("\n Example 3 --------------------------- \n")
+
+housing = fetch_california_housing()
+n_samples = 500
+X = housing.data[:n_samples]
+y = housing.target[:n_samples]
+# split data into training test and test set
+np.random.seed(15029)
+X_train, X_test, y_train, y_test = train_test_split(X, y, 
+                                                    test_size=0.2)
+obj = ms.GenericBoostingRegressor(regr, n_hidden_features=2, n_estimators=10)
+
+
+obj.fit(X_train, y_train)
+print(f"obj.obj['fit_obj_i'][0]: {obj.obj['fit_obj_i'][0].coef_}")
+print(f"score: {np.sqrt(np.mean(np.square(obj.predict(X_test[3:10,:]) - y_test[3:10])))}")
+
+obj = obj.update(X_test[0,:], y_test[0])
+obj = obj.update(X_test[1,:], y_test[1])
+obj = obj.update(X_test[2,:], y_test[2])
+print(f"obj.obj['fit_obj_i'][0]: {obj.obj['fit_obj_i'][0].coef_}")
+print(f"preds: {obj.predict(X_test[3:10,:])}")
+print(f"score: {np.sqrt(np.mean(np.square(obj.predict(X_test[3:10,:]) - y_test[3:10])))}")

examples/lazy_booster_regression.py

@@ -18,8 +18,8 @@
 print(models)
 
 data = fetch_california_housing()
-X = data.data
-y= data.target
+X = data.data[0:1000,:]
+y= data.target[0:1000]
 X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = .2, random_state = 123)
 
 regr = ms.LazyBoostingRegressor(verbose=0, ignore_warnings=True,