Updated Results k=5 folds

Author: jmarrietar

Algorithms/MILBoost.py

@@ -103,4 +103,4 @@ def predict(self,test_bags):
         for i in range (0,B-1):
             out[i]=1-np.prod([1-np.asarray(pij[IbagT[i]])])
 
-        return out
+        return out 

data/.DS_Store

@@ -14,6 +14,8 @@
 from sklearn import cross_validation
 from sklearn import metrics
 from data import load_data
+from sklearn.utils import shuffle
+import random as rand
 
 #Import Algorithms 
 from MILpy.Algorithms.simpleMIL import simpleMIL
@@ -29,23 +31,28 @@
 """
 
 #Load Data 
-bags,labels,X = load_data('musk1_scaled')  #Musk1 Escalado
-#bags,labels,X = load_data('musk1_original')  #Musk1 Original  ALGO PASA CON ESTA EN EMDD Y MAXDD
-#bags,labels,X = load_data('data_gauss')  #Gaussian data
-#bags,labels,X = load_data('fox_original')  #Fox Original
-#bags,labels,X = load_data('fox_scaled')    #Fox Escalado
+#bags,labels,_ = load_data('musk1_scaled')  #Musk1 Escalado
+#bags,labels,_ = load_data('musk1_original')  #Musk1 Original  ALGO PASA CON ESTA EN EMDD Y MAXDD
+bags,labels,_ = load_data('data_gauss')  #Gaussian data
+#bags,labels,_ = load_data('fox_original')  #Fox Original
+#bags,labels,_ = load_data('fox_scaled')    #Fox Escalado
 
 seed = 66
 #seed = 70
 #Split Data
 #seed= 90
+
+#Shuffle Data
+bags,labels = shuffle(bags, labels, random_state=rand.randint(0, 100))
+
 train_bags, test_bags, train_labels, test_labels = cross_validation.train_test_split(bags, labels, test_size=0.1, random_state=seed)
 
                             ################
                             #Bags Of Words #
                             ################
 bow_classifier = BOW() 
-bow_classifier.fit(train_bags, train_labels,k=100,covar_type = 'diag',n_iter = 20)
+#bow_classifier.fit(train_bags, train_labels,k=100,covar_type = 'diag',n_iter = 20)
+bow_classifier.fit(train_bags, train_labels,k=10,covar_type = 'diag',n_iter = 20)
 predictions = bow_classifier.predict(test_bags)
 accuracie = np.average(test_labels.T == np.sign(predictions))
 print '\n Accuracy: %.2f%%' % (100 * accuracie)
@@ -135,7 +142,7 @@
                             ##########
 #Nota Importante:  Solo Funciona Con musk1 original. 
 #Load Data 
-bags,labels,X = load_data('musk1_original')  #Musk1 Original
+bags,labels,_ = load_data('musk1_original')  #Musk1 Original
 seed = 90
 train_bags, test_bags, train_labels, test_labels = cross_validation.train_test_split(bags, labels, test_size=0.1, random_state=seed)
 
@@ -148,7 +155,7 @@
                             #######
                             #MILES#
                             #######
-bags,labels,X = load_data('data_gauss')  #Gaussian data
+bags,labels,_ = load_data('data_gauss')  #Gaussian data
 seed = 66
 train_bags, test_bags, train_labels, test_labels = cross_validation.train_test_split(bags, labels, test_size=0.1, random_state=seed)
 

data/musk1_py/bags.pckl

@@ -11,6 +11,8 @@
 import sys,os
 os.chdir('/Users/josemiguelarrieta/Documents/MILpy')
 sys.path.append(os.path.realpath('..'))
+from sklearn.utils import shuffle
+import random as rand
 from data import load_data
 from MILpy.functions.mil_cross_val import mil_cross_val
 
@@ -20,64 +22,68 @@
 from MILpy.Algorithms.maxDD import maxDD
 from MILpy.Algorithms.CKNN import CKNN
 from MILpy.Algorithms.EMDD import EMDD
-from MILpy.Algorithms.MILES import MILES
 from MILpy.Algorithms.BOW import BOW
+from MILpy.Algorithms.MILES import MILES
 
 """
 Note: There is an Issue with regars musk1_original and EMDD and maxDD.
 """
 
 #Load Data 
-#bags,labels,X = load_data('musk1_scaled')
-#bags,labels,X = load_data('musk2_scaled') # No finish. 
+#bags,labels,_ = load_data('musk1_scaled')
+#bags,labels,_ = load_data('musk2_scaled') 
+#bags,labels,_ = load_data('fox_scaled') 
+#bags,labels,_ = load_data('tiger_scaled')  
+#bags,labels,_ = load_data('elephant_scaled') 
+bags,labels,_ = load_data('data_gauss')
 
-bags,labels,X = load_data('fox_scaled') # No finish. 
-bags,labels,X = load_data('tiger_scaled') # No finish. 
-bags,labels,X = load_data('elephant_scaled') # No finish. 
 
+#Shuffle Data
+bags,labels = shuffle(bags, labels, random_state=rand.randint(0, 100))
 
+#Number of Folds 
+folds=5
 
 bow_classifier = BOW() 
-parameters_bow = {'k':100,'covar_type':'diag','n_iter':20}
-accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=bow_classifier, folds=10,parameters=parameters_bow)
+#parameters_bow = {'k':100,'covar_type':'diag','n_iter':20}
+parameters_bow = {'k':10,'covar_type':'diag','n_iter':20}
+accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=bow_classifier, folds=folds, parameters=parameters_bow)
 
 SMILa = simpleMIL()
 parameters_smil = {'type': 'max'}
 #En este me funciono maxDD porque no tiene problem con parametros 
-accuracie, results_accuracie, auc,results_auc,elapsed  = mil_cross_val(bags=bags,labels=labels, model=SMILa, folds=10,parameters=parameters_smil,timer=True)
+accuracie, results_accuracie, auc,results_auc,elapsed  = mil_cross_val(bags=bags,labels=labels, model=SMILa, folds=folds, parameters=parameters_smil,timer=True)
 
 parameters_smil = {'type': 'min'}
 #En este me funciono maxDD porque no tiene problem con parametros 
-accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=SMILa, folds=10,parameters=parameters_smil)
+accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=SMILa, folds=folds, parameters=parameters_smil)
 
 parameters_smil = {'type': 'extreme'}
 #En este me funciono maxDD porque no tiene problem con parametros 
-accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=SMILa, folds=10,parameters=parameters_smil)
+accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=SMILa, folds=folds, parameters=parameters_smil)
 
 parameters_smil = {'type': 'average'}
 #En este me funciono maxDD porque no tiene problem con parametros 
-accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=SMILa, folds=10,parameters=parameters_smil)
+accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=SMILa, folds=folds, parameters=parameters_smil)
 
 cknn_classifier = CKNN() 
 parameters_cknn = {'references': 3, 'citers': 5}
-accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=cknn_classifier, folds=10,parameters=parameters_cknn)
+accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=cknn_classifier, folds=folds, parameters=parameters_cknn)
 
 maxDD_classifier = maxDD()
-accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=maxDD_classifier, folds=10,parameters={})
+accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=maxDD_classifier, folds=folds, parameters={})
 
 emdd_classifier = EMDD()
-accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=emdd_classifier, folds=10,parameters={})
+accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=emdd_classifier, folds=folds, parameters={})
 
-
-#STOP AQUIIIIII ARREGLAR ESTO 
+#STOP: Fix This. 
 #MIL BOOST: Tiene Out y no tiene predicted. 
 
-out = []
-for i in range(0,1):
-    milboost_classifier = MILBoost() 
-    accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=milboost_classifier, folds=10,parameters={})
-    out.append(auc)
+milboost_classifier = MILBoost() 
+accuracie, results_accuracie, auc,results_auc  = mil_cross_val(bags=bags,labels=labels, model=milboost_classifier, folds=10,parameters={})
+
 
 #MIL BOOST: Tiene Out y no tiene predicted. 
 
+#Stop Here: 
 #MILES: Tiene Out no tiene predicted.