def CrossValidationWithSampling():
runObj = run_util.run()
## Load data
((X_train,Y_train),(X_test)) = loadData(runObj)
## Sample train data
RSObj=RS.randomSampling()
(X_train,Y_train) = RSObj.getRandomSample(X_train,Y_train)
(X_train,Y_train) = runObj.randomShuffleData(X_train,Y_train)
print('\n Size of X_train : {0}').format(X_train.shape)
print('\n Size of Y_train : {0}').format(Y_train.shape)
X_train.reset_index(drop=True,inplace=True)
Y_train.reset_index(drop=True,inplace=True)
## Cross Validation
kf = KFold(X_train.shape[0], n_folds=5)
for train_index, test_index in kf:
X_train_CV, X_test_CV = X_train.ix[train_index], X_train.ix[test_index]
Y_train_CV, Y_test_CV = Y_train.ix[train_index], Y_train.ix[test_index]
print('\n Size of X_train_CV : {0}').format(X_train_CV.shape)
print('\n Size of Y_train_CV : {0}').format(Y_train_CV.shape)
print('\n Size of X_test_CV : {0}').format(X_test_CV.shape)
print('\n Size of Y_test_CV : {0}').format(Y_test_CV.shape)
## Run classifiers
runObj.runClassifier(X_train_CV,Y_train_CV,X_test_CV,Y_test_CV)
def RFEWithSampling():
runObj = run_util.run()
## Load data
((X_train,Y_train),(X_test)) = loadData(runObj)
## Sample train data
RSObj=RS.randomSampling()
(X_train,Y_train)=RSObj.getRandomSample(X_train,Y_train)
(X_train,Y_train) = runObj.randomShuffleData(X_train,Y_train)
print('\n Size of Sample X_train : {0}').format(X_train.shape)
print('\n Size of Sample Y_train : {0}').format(Y_train.shape)
X_train.reset_index(drop=True,inplace=True)
Y_train.reset_index(drop=True,inplace=True)
## RFE Feature Selection
print('\n RFE Feature Selection starts...')
selected_columns = AF.RFE_featureSelection(X_train,Y_train)
print('\n RFE Feature Selection ends...')
X_train = X_train[selected_columns]
X_test = X_train[selected_columns]
## Split data into 66% train and 33% test
((X_train_S,Y_train_S), (X_test_S,Y_test_S)) = runObj.splitDataset(X_train,Y_train)
print('\n Size of X_train_S : {0}').format(X_train_S.shape)
print('\n Size of Y_train_S : {0}').format(Y_train_S.shape)
print('\n Size of X_test_S : {0}').format(X_test_S.shape)
print('\n Size of Y_test_S : {0}').format(Y_test_S.shape)
## Run classifiers
runObj.runClassifier(X_train_S,Y_train_S,X_test_S,Y_test_S)
def costValuesBasedPrediction():
runObj = run_util.run()
## Load data
((X_train,Y_train),(X_test)) = loadData(runObj)
## Select feature related to cost
costFeatures = ['MMRAcquisitionAuctionAveragePrice','VehBCost']
X_train = X_train[costFeatures]
print('\n Size of X_train only cost features : {0}').format(X_train.shape)
## Sample train data
RSObj=RS.randomSampling()
(X_train,Y_train) = RSObj.getRandomSample(X_train,Y_train)
(X_train,Y_train) = runObj.randomShuffleData(X_train,Y_train)
print('\n Size of X_train : {0}').format(X_train.shape)
print('\n Size of Y_train : {0}').format(Y_train.shape)
X_train.reset_index(drop=True,inplace=True)
Y_train.reset_index(drop=True,inplace=True)
## Split data into 66% train and 33% test
((X_train_S,Y_train_S), (X_test_S,Y_test_S)) = runObj.splitDataset(X_train,Y_train)
print('\n Size of X_train_S : {0}').format(X_train_S.shape)
print('\n Size of Y_train_S : {0}').format(Y_train_S.shape)
print('\n Size of X_test_S : {0}').format(X_test_S.shape)
print('\n Size of Y_test_S : {0}').format(Y_test_S.shape)
## Run classifiers
runObj.runClassifier(X_train_S,Y_train_S,X_test_S,Y_test_S)
def RFE_featureSelection(X_train,Y_train):
## Sampling
RSObj=randomSampling.randomSampling()
(X_train,Y_train)=RSObj.getRandomSample(X_train,Y_train)
X_train.reset_index(drop=True,inplace=True)
Y_train.reset_index(drop=True,inplace=True)
## Select classifier and parameters
logistic = linear_model.LogisticRegression(C=10, class_weight=None, dual=False, fit_intercept=True,
intercept_scaling=1, max_iter=100, multi_class='ovr', n_jobs=1,
penalty='l1', random_state=None, solver='liblinear', tol=0.01,
verbose=0, warm_start=False)
## Initialiaze RFE
rfecv = RFECV(estimator=logistic, step=1, cv=5,
scoring='recall')
## Fit data
rfecv.fit(X_train, Y_train)
## Selected Features
print("Optimal number of features : %d" % rfecv.n_features_)
## Plot importance
plt.figure()
plt.xlabel("Number of features selected")
plt.ylabel("Cross validation score")
plt.plot(range(1, len(rfecv.grid_scores_) + 1), rfecv.grid_scores_)
plt.show()
#print('\n Selectd Columns : {0}').format(list(rfecv.support_))
print('\n Selectd Columns : {0}').format(X_train.columns[list(rfecv.support_)])
selected_columns = X_train.columns[list(rfecv.support_)]
return selected_columns
def RFE_featureSelection(X_train, Y_train):
## Sampling
RSObj = randomSampling.randomSampling()
(X_train, Y_train) = RSObj.getRandomSample(X_train, Y_train)
X_train.reset_index(drop=True, inplace=True)
Y_train.reset_index(drop=True, inplace=True)
## Select classifier and parameters
logistic = linear_model.LogisticRegression(C=10,
class_weight=None,
dual=False,
fit_intercept=True,
intercept_scaling=1,
max_iter=100,
multi_class='ovr',
n_jobs=1,
penalty='l1',
random_state=None,
solver='liblinear',
tol=0.01,
verbose=0,
warm_start=False)
## Initialiaze RFE
rfecv = RFECV(estimator=logistic, step=1, cv=5, scoring='recall')
## Fit data
rfecv.fit(X_train, Y_train)
## Selected Features
print("Optimal number of features : %d" % rfecv.n_features_)
## Plot importance
plt.figure()
plt.xlabel("Number of features selected")
plt.ylabel("Cross validation score")
plt.plot(range(1, len(rfecv.grid_scores_) + 1), rfecv.grid_scores_)
plt.show()
#print('\n Selectd Columns : {0}').format(list(rfecv.support_))
print('\n Selectd Columns : {0}').format(X_train.columns[list(
rfecv.support_)])
selected_columns = X_train.columns[list(rfecv.support_)]
return selected_columns
def Boosting_featureSelection(X_train, Y_train):
## Feature selection based on GradientBoostedTrees
## Sampling
RSObj = randomSampling.randomSampling()
(X_train, Y_train) = RSObj.getRandomSample(X_train, Y_train)
X_train.reset_index(drop=True, inplace=True)
Y_train.reset_index(drop=True, inplace=True)
## Build a forest and compute the feature importances
forest = ExtraTreesClassifier(n_estimators=100)
## Fit Forest
forest.fit(X_train, Y_train)
importances = forest.feature_importances_
std = np.std([tree.feature_importances_ for tree in forest.estimators_],
axis=0)
indices = np.argsort(importances)[::-1]
## Print the feature ranking
print("Feature ranking:")
cols = list(X_train.columns)
for f in range(X_train.shape[1]):
print("%d. feature %s (%f)" %
(f + 1, cols[indices[f]], importances[indices[f]]))
## Plot the feature importances of the forest
plt.figure()
plt.title("Feature importances")
plt.bar(range(X_train.shape[1]),
importances[indices],
color="r",
yerr=std[indices],
align="center")
plt.xticks(range(X_train.shape[1]), [cols[i] for i in indices])
plt.xlim([-1, X_train.shape[1]])
plt.show()
## Generate dictionary of column importance
cols = [cols[i] for i in indices]
dictionary = dict(zip(cols, importances))
return dictionary
def Boosting_featureSelection(X_train,Y_train):
## Feature selection based on GradientBoostedTrees
## Sampling
RSObj=randomSampling.randomSampling()
(X_train,Y_train)=RSObj.getRandomSample(X_train,Y_train)
X_train.reset_index(drop=True,inplace=True)
Y_train.reset_index(drop=True,inplace=True)
## Build a forest and compute the feature importances
forest = ExtraTreesClassifier(n_estimators=100)
## Fit Forest
forest.fit(X_train, Y_train)
importances = forest.feature_importances_
std = np.std([tree.feature_importances_ for tree in forest.estimators_],
axis=0)
indices = np.argsort(importances)[::-1]
## Print the feature ranking
print("Feature ranking:")
cols = list(X_train.columns)
for f in range(X_train.shape[1]):
print("%d. feature %s (%f)" % (f + 1, cols[indices[f]], importances[indices[f]]))
## Plot the feature importances of the forest
plt.figure()
plt.title("Feature importances")
plt.bar(range(X_train.shape[1]), importances[indices],
color="r", yerr=std[indices], align="center")
plt.xticks(range(X_train.shape[1]), [cols[i] for i in indices])
plt.xlim([-1, X_train.shape[1]])
plt.show()
## Generate dictionary of column importance
cols = [cols[i] for i in indices]
dictionary = dict(zip(cols,importances))
return dictionary
def BoostingWithSampling():
runObj = run_util.run()
## Load data
((X_train,Y_train),(X_test)) = loadData(runObj)
## Sample train data
RSObj=RS.randomSampling()
(X_train,Y_train)=RSObj.getRandomSample(X_train,Y_train)
(X_train,Y_train) = runObj.randomShuffleData(X_train,Y_train)
print('\n Size of X_train : {0}').format(X_train.shape)
print('\n Size of Y_train : {0}').format(Y_train.shape)
X_train.reset_index(drop=True,inplace=True)
Y_train.reset_index(drop=True,inplace=True)
## Boosting Feature Selection
print('\n Boosting Feature Selection starts...')
selected_columns_dict = AF.Boosting_featureSelection(X_train,Y_train)
print('\n Boosting Feature Selection ends...')
selected_columns=[]
for col, imp in selected_columns_dict.iteritems():
if(imp > 0.005):
selected_columns.append(col)
print('\n Selected Cols : {0}').format(len(selected_columns))
X_train = X_train[selected_columns]
X_test = X_train[selected_columns]
## Split data into 66% train and 33% test
((X_train_S,Y_train_S), (X_test_S,Y_test_S)) = runObj.splitDataset(X_train,Y_train)
print('\n Size of X_train_S : {0}').format(X_train_S.shape)
print('\n Size of Y_train_S : {0}').format(Y_train_S.shape)
print('\n Size of X_test_S : {0}').format(X_test_S.shape)
print('\n Size of Y_test_S : {0}').format(Y_test_S.shape)
## Run classifiers
runObj.runClassifier(X_train_S,Y_train_S,X_test_S,Y_test_S)
def handpickWithSampling():
runObj = run_util.run()
## Load data
((X_train,Y_train),(X_test)) = loadData(runObj)
## Sample train data
RSObj=RS.randomSampling()
(X_train,Y_train) = RSObj.getRandomSample(X_train,Y_train)
(X_train,Y_train) = runObj.randomShuffleData(X_train,Y_train)
print('\n Size of X_train : {0}').format(X_train.shape)
print('\n Size of Y_train : {0}').format(Y_train.shape)
X_train.reset_index(drop=True,inplace=True)
Y_train.reset_index(drop=True,inplace=True)
## Split data into 66% train and 33% test
((X_train_S,Y_train_S), (X_test_S,Y_test_S)) = runObj.splitDataset(X_train,Y_train)
print('\n Size of X_train_S : {0}').format(X_train_S.shape)
print('\n Size of Y_train_S : {0}').format(Y_train_S.shape)
print('\n Size of X_test_S : {0}').format(X_test_S.shape)
print('\n Size of Y_test_S : {0}').format(Y_test_S.shape)
## Run classifiers
runObj.runClassifier(X_train_S,Y_train_S,X_test_S,Y_test_S)
print('\n predictions : {0}').format(predictions.shape[0])
for i in range(RefId.shape[0]):
data.append([RefId[i],predictions[i]])
print('\n Data : \n {0}').format(data)
with open(fileName, 'wb') as fp:
a = csv.writer(fp, delimiter=',')
a.writerows(data)
if __name__=="__main__":
fileName='updatedTraining.csv'
dataLoadObj=dataLoad()
Traindataset=dataLoadObj.loadData(fileName,'train')
############ Random Sampling ###################
RSObj=RS.randomSampling()
Traindataset=RSObj.getRandomSample(Traindataset)
print('\n Train : {0}').format(Traindataset.shape)
'''
fileName='updatetest.csv'
testDataset=dataLoadObj.loadData(fileName,'test')
testDataset=testDataset[0:-6]
print('\n Test : {0}').format(testDataset.shape)
dataLoadObj.runClassifiers(Traindataset,testDataset)
'''
print('\n Cross Validation : ')
randindices = np.random.randint(0,Traindataset.shape[0],10000)
Traindataset=Traindataset[randindices,:]
Kobj=Kfold.Kfold(No_of_Folds=10)
kf=Kobj.getFoldIndices(Traindataset)
