def classify():
linearSVM= LinearSVC( random_state=666, class_weight="balanced", max_iter=5000, C=2.0,tol=0.001, dual=True )
linearSVM_SVC= SVC( C=1, kernel="rbf", tol=1, random_state=0,gamma=1 )
logistic = LogisticRegression( fit_intercept=True,class_weight="balanced", n_jobs=-1, C=1.0,
max_iter=200 )
rand_forest = RandomForestClassifier( n_estimators=403, random_state=666, max_depth=73, n_jobs=-1 )
bc=BaggingClassifier( base_estimator=logistic, n_estimators=403, n_jobs=-1, random_state=666,
max_features=410)
ensemble_voting=VotingClassifier([("logistic",logistic),("rand_forest",rand_forest),("sgdc",SGDClassifier())],weights=[1,1,2])
boost = AdaBoostClassifier(base_estimator=logistic)
xgboost= XGBoostClassifier( n_estimators=103, seed=666, max_depth=4, objective="multi:softmax" )
return ensemble_voting
def fit(self, X, y, **fit_params):
result = XGBClassifierImpl(self.max_depth, self.learning_rate, self.n_estimators,
self.verbosity, self.objective, self.booster, self.n_jobs,
self.nthread, self.gamma, self.min_child_weight, self.max_delta_step, self.subsample,
self.colsample_bytree, self.colsample_bylevel, self.colsample_bynode, self.reg_alpha,
self.reg_lambda, self.scale_pos_weight, self.base_score, self.random_state,
self.seed, self.missing)
result._xgboost_model = XGBoostClassifier(
**self.get_params())
if fit_params is None:
result._xgboost_model.fit(X, y)
else:
result._xgboost_model.fit(X, y, **fit_params)
return result
rf_acc = cv(RandomForestClassifier(n_estimators=403,n_jobs=-1, random_state=seed),data_model.drop(columns='label',axis=1), data_model['label']) # ## Experiment 3: Added Feature + XGBoost # In[55]: from xgboost import XGBClassifier as XGBoostClassifier # In[56]: X_train, X_test, y_train, y_test = train_test_split(data_model.drop(columns='label',axis=1),data_model['label'] , test_size=0.3) precision, recall, accuracy, f1 = test_classifier(X_train, y_train, X_test, y_test, XGBoostClassifier(seed=seed)) # ## Experiment 4: Added Feature + Naive Bayes # In[57]: X_train, X_test, y_train, y_test = train_test_split(data_model.drop(columns='label',axis=1),data_model['label'] , test_size=0.3) precision, recall, accuracy, f1 = test_classifier(X_train, y_train, X_test, y_test, BernoulliNB()) # ## Test Data # In[58]:
def best_fit(X_train, y_train):
log("")
seed = 666
import time as ttt
attributes = len(X_train.columns)
examples = len(X_train)
now = time()
log(ttt.ctime())
# Parameters for SVM
# parameters = {
# "dual": [True, False],
# "tol": [1e-3, 1e-4, 1e-5],
# "C": [1.0, 1.5, 2.0, 5.0, 10, 100, 1000]
# }
# rand_search = RandomizedSearchCV(LinearSVC(max_iter=5000), param_distributions=parameters, cv=8,n_jobs=-1,n_iter=20)
#
#
# rand_search.fit(X_train,y_train)
# report(rand_search.cv_results_, 10)
# log(ttt.ctime())
# log(time() - now)
# return
# Parameters for Bagging
# parameters = {
# "n_estimators": [2, 3, 5, 13, 51, 201, 303, 403, 505],
# "max_features": list(map(lambda x: int(x),
# [sqrt(attributes), 2 * sqrt(attributes), 3 * sqrt(attributes), attributes / 2,
# attributes / 3, attributes / 4]))
# }
#
# rand_search = RandomizedSearchCV(BaggingClassifier(
# base_estimator=LinearSVC(random_state=seed, class_weight="balanced", max_iter=5000, C=1.0, tol=0.0001, dual=True),
# random_state=seed, n_jobs=1), param_distributions=parameters, n_jobs=-1, n_iter=3, cv=8,
# scoring=make_scorer(f1_score, average="micro", labels=["positive", "negative", "neutral"]))
#
# now = time()
# log(ttt.ctime())
# rand_search.fit(X_train, y_train)
#
# report(rand_search.cv_results_, 10)
log(ttt.ctime())
log(time() - now)
# Parameters for RF
# log("RF:")
# parameters = {
# "n_estimators":[103, 201, 305, 403, 666, 1001, 5007, 10001],
# "max_depth":[None, 5, 20, 40, 73, 100, 1000, 2000],
# "criterion":["gini", "entropy"]
# }
#
# rand_search = RandomizedSearchCV(RandomForestClassifier(random_state=seed,n_jobs=-1),param_distributions=parameters,
# n_iter=15,scoring="accuracy",
# n_jobs=1,cv=10)
# now = time()
# log(ttt.ctime())
# rand_search.fit(X_train, y_train)
#
# report(rand_search.cv_results_, 10)
# log(ttt.ctime())
# log(time() - now)
# Parameters for XGBoost
log("XGB:")
parameters = {
"n_estimators": [103, 201, 403],
"max_depth": [3, 10, 15],
"objective": ["multi:softmax", "binary:logistic"],
"learning_rate": [0.05, 0.1, 0.15, 0.3]
}
rand_search = RandomizedSearchCV(XGBoostClassifier(seed=seed),
param_distributions=parameters,
n_iter=5,
scoring="accuracy",
n_jobs=-1,
cv=8)
now = time()
log(ttt.ctime())
rand_search.fit(X_train, y_train)
report(rand_search.cv_results_, 10)
log(ttt.ctime())
log(time() - now)
parameters = {
"n_estimators": [403, 666, 1000],
"max_depth": [40, 50, 90, 100, 200],
"subsample": [1.0, 0.6, 0.9],
"objective": ["multi:softmax", "binary:logistic"],
"learning_rate": [0.1, 0.15, 0.5]
}
rand_search = RandomizedSearchCV(XGBoostClassifier(seed=seed, ),
param_distributions=parameters,
n_iter=5,
scoring="accuracy",
n_jobs=-1,
cv=8)
now = time()
log(ttt.ctime())
rand_search.fit(X_train, y_train)
report(rand_search.cv_results_, 10)
log(ttt.ctime())
log(time() - now)
return
# Parameters for VotingClassifier
# parameters = {
# "weights": [
# [1, 1, 1],
# [2, 1, 1],
# [2, 2, 1],
# [4, 1, 5],
# [1, 1, 2],
# [5, 1, 2],
# [5, 2, 1],
# [5, 3, 2],
# [6, 2, 1],
# [6, 1, 5],
# [6, 1, 2],
# [7, 1, 6],
# [7, 2, 3],
# ]
# }
log("Voting RF XGB NB:")
parameters = {
"weights": [[1, 1, 1], [2, 1, 1], [1, 1, 2], [4, 1, 5], [3, 1, 3],
[3, 1, 4]]
}
rand_search = GridSearchCV(VotingClassifier(
[("randomforest",
RandomForestClassifier(
n_estimators=403, random_state=seed, max_depth=73, n_jobs=-1)),
("naivebayes", BernoulliNB()),
("xgboost",
XGBoostClassifier(n_estimators=103,
seed=seed,
max_depth=3,
objective="multi:softmax"))],
voting="soft",
n_jobs=1),
scoring="accuracy",
n_jobs=-1,
cv=8,
param_grid=parameters)
rand_search.fit(X_train, y_train)
#
report(rand_search.cv_results_, 10)
log(ttt.ctime())
log(time() - now)
from xgboost import XGBClassifier as XGBoostClassifier
X_train, X_test, y_train, y_test = train_test_split(data_model.drop(
columns='label', axis=1),
data_model['label'],
test_size=0.3)
precision, recall, accuracy, f1 = test_classifier(X_train, y_train, X_test,
y_test,
XGBoostClassifier(seed=seed))
C=1.0,
max_iter=200)
rand_forest = RandomForestClassifier(n_estimators=403,
random_state=666,
max_depth=73,
n_jobs=-1)
bc = BaggingClassifier(base_estimator=logistic,
n_estimators=403,
n_jobs=-1,
random_state=666,
max_features=410)
ensemble_voting = VotingClassifier([("svm", linearSVM_SVC),
("logistic", logistic),
("rand_forest", rand_forest),
("sgdc", SGDClassifier())],
weights=[1, 2, 1, 1])
boost = AdaBoostClassifier(base_estimator=logistic)
xgboost = XGBoostClassifier(n_estimators=103,
seed=666,
max_depth=4,
objective="multi:softmax")
X_train, X_test, Y_train, Y_test = train_test_split(left, right, test_size=0.1)
text_clf = Pipeline([('vect', CountVectorizer()),
('tfidf', TfidfTransformer()),
('clf-svm', ensemble_voting)])
text_clf = text_clf.fit(X_train, Y_train)
predicted = text_clf.predict(X_test)
print classification_report(Y_test, predicted)
cdTest = dataCleanup()
cdTest.init(data_dir + 'tweets.csv', isTestingSet=True)
cdTest.buildFeatures()
cdTest.cleanData()
tTest = cdTest.processedData
stTest = stemAndTokenizeData()
tTest = stTest.tokenize(tTest)
tTest = stTest.stem(tTest)
bwTest = buildWordList()
bwTest.buildWordListFunction(tTest)
bowTest = bagOfWords()
bowTest.buildDataModel(tTest, bwTest.wordList, uW2V, isTestingSet=True)
dataModelTest = bowTest.dataModel
print("Testing Model built!")
xgboost = XGBoostClassifier(seed=seed,
n_estimators=403,
max_depth=10,
objective="binary:logistic",
learning_rate=0.15)
xgboost.fit(dataModel.iloc[:, 1:], dataModel.iloc[:, 0])
print("Training Finished!")
predictions = xgboost.predict(dataModelTest.iloc[:, 1:])
results = pd.DataFrame([], columns=["Id", "Category"])
results["Id"] = dataModelTest["original_id"].astype("int64")
results["Category"] = predictions
results.to_csv("results.csv", index=False)
print("Results have been saved to file!!")
import pandas as pd
from sklearn.preprocessing import LabelEncoder
from sklearn.model_selection import train_test_split
from xgboost import XGBoostClassifier
from sklearn.metrics import accuracy_score
data = pd.read_csv('Loan payments data.csv')
print(data.head())
X = data.iloc['Principal','terms','age','education','Gender']
Y=data['loan_status']
gender_label = LabelEncoder()
gender_label.fit(X.Gender)
X['Gender_Labelled']=gender_label.transform(X.Gender)
Ed_Label = LabelEncoder()
Ed_Label.fit(X.education)
X['Ed_Labelled'] = Ed_Label.transform(X.education)
Y_label = LabelEncoder
Y_label.fit(Y)
Y_labelled = Y_label.transform(Y)
X_Labelled = X.iloc('Principal','terms','age','Ed_Labelled','Gender_Labelled')
x_train,x_test,y_train,y_test = train_test_split(X_Labelled,Y_labelled)
model = XGBoostClassifier()
model.fit(x_train,y_train)
y_predict = model.predict(x_test)
print(accuracy_score(y_test,y_predict))
def best_fit(X_train, y_train):
log("")
seed = 666
import time as ttt
attributes = len(X_train.columns)
examples = len(X_train)
now = time()
log(ttt.ctime())
log(ttt.ctime())
log(time() - now)
log("XGB:")
parameters = {
"n_estimators": [103, 201, 403],
"max_depth": [3, 10, 15],
"objective": ["multi:softmax", "binary:logistic"],
"learning_rate": [0.05, 0.1, 0.15, 0.3]
}
rand_search = RandomizedSearchCV(XGBoostClassifier(seed=seed),
param_distributions=parameters,
n_iter=5,
scoring="accuracy",
n_jobs=-1,
cv=8)
now = time()
log(ttt.ctime())
rand_search.fit(X_train, y_train)
report(rand_search.cv_results_, 10)
log(ttt.ctime())
log(time() - now)
parameters = {
"n_estimators": [403, 666, 1000],
"max_depth": [40, 50, 90, 100, 200],
"subsample": [1.0, 0.6, 0.9],
"objective": ["multi:softmax", "binary:logistic"],
"learning_rate": [0.1, 0.15, 0.5]
}
rand_search = RandomizedSearchCV(XGBoostClassifier(seed=seed, ),
param_distributions=parameters,
n_iter=5,
scoring="accuracy",
n_jobs=-1,
cv=8)
now = time()
log(ttt.ctime())
rand_search.fit(X_train, y_train)
report(rand_search.cv_results_, 10)
log(ttt.ctime())
log(time() - now)
return
log("Voting RF XGB NB:")
parameters = {
"weights": [[1, 1, 1], [2, 1, 1], [1, 1, 2], [4, 1, 5], [3, 1, 3],
[3, 1, 4]]
}
rand_search = GridSearchCV(VotingClassifier(
[("randomforest",
RandomForestClassifier(
n_estimators=403, random_state=seed, max_depth=73, n_jobs=-1)),
("naivebayes", BernoulliNB()),
("xgboost",
XGBoostClassifier(n_estimators=103,
seed=seed,
max_depth=3,
objective="multi:softmax"))],
voting="soft",
n_jobs=1),
scoring="accuracy",
n_jobs=-1,
cv=8,
param_grid=parameters)
rand_search.fit(X_train, y_train)
#
report(rand_search.cv_results_, 10)
log(ttt.ctime())
log(time() - now)