kernel_id
int64
24.2k
23.3M
prompt
stringlengths
8
1.85M
completetion
stringlengths
1
182k
comp_name
stringlengths
5
57
9,251,987
train_df = pd.read_csv('/kaggle/input/nlp-getting-started/train.csv') test_df = pd.read_csv('/kaggle/input/nlp-getting-started/test.csv' )<prepare_x_and_y>
raw_models = model_check(X, y, estimators, cv) display(raw_models.style.background_gradient(cmap='summer_r'))
Titanic - Machine Learning from Disaster
9,251,987
X = train_df.loc[:,'text'] y = train_df.loc[:,'target']<define_variables>
def m_roc(estimators, cv, X, y): fig, axes = plt.subplots(math.ceil(len(estimators)/ 2), 2, figsize=(25, 50)) axes = axes.flatten() for ax, estimator in zip(axes, estimators): tprs = [] aucs = [] mean_fpr = np.linspace(0, 1, 100) for i,(train, test)in enumerate(cv.split(X, y)) : estimator.fit(X.loc[train], y.loc[train]) viz = plot_roc_curve(estimator, X.loc[test], y.loc[test], name='ROC fold {}'.format(i), alpha=0.3, lw=1, ax=ax) interp_tpr = interp(mean_fpr, viz.fpr, viz.tpr) interp_tpr[0] = 0.0 tprs.append(interp_tpr) aucs.append(viz.roc_auc) ax.plot([0, 1], [0, 1], linestyle='--', lw=2, color='r', label='Chance', alpha=.8) mean_tpr = np.mean(tprs, axis=0) mean_tpr[-1] = 1.0 mean_auc = auc(mean_fpr, mean_tpr) std_auc = np.std(aucs) ax.plot(mean_fpr, mean_tpr, color='b', label=r'Mean ROC(AUC = %0.2f $\pm$ %0.2f)' % (mean_auc, std_auc), lw=2, alpha=.8) std_tpr = np.std(tprs, axis=0) tprs_upper = np.minimum(mean_tpr + std_tpr, 1) tprs_lower = np.maximum(mean_tpr - std_tpr, 0) ax.fill_between(mean_fpr, tprs_lower, tprs_upper, color='grey', alpha=.2, label=r'$\pm$ 1 std.dev.') ax.set(xlim=[-0.02, 1.02], ylim=[-0.02, 1.02], title=f'{estimator.__class__.__name__} ROC') ax.legend(loc='lower right', prop={'size': 18}) plt.show()
Titanic - Machine Learning from Disaster
9,251,987
max_len = 0 for text in X: max_len = max(max_len, len(text)) max_len<prepare_x_and_y>
m_roc(estimators, cv, X, y )
Titanic - Machine Learning from Disaster
9,251,987
class Dataset(torch.utils.data.Dataset): def __init__(self,df,y=None,max_len=164): self.df = df self.y = y self.max_len= max_len self.tokenizer = transformers.RobertaTokenizer.from_pretrained('roberta-base') def __getitem__(self,index): row = self.df.iloc[index] ids,masks = self.get_input_data(row) data = {} data['ids'] = ids data['masks'] = masks if self.y is not None: data['out'] = torch.tensor(self.y.iloc[index],dtype=torch.float32) return data def __len__(self): return len(self.df) def get_input_data(self,row): row = self.tokenizer.encode(row,add_special_tokens=True,add_prefix_space=True) padded = row + [0]*(max_len - len(row)) padded = torch.tensor(padded,dtype=torch.int64) mask = torch.where(padded != 0 , torch.tensor(1),torch.tensor(0)) return padded,mask <split>
f_imp(estimators, X, y, 14 )
Titanic - Machine Learning from Disaster
9,251,987
train_x,val_x,train_y,val_y = train_test_split(X,y,test_size=0.2,stratify=y) train_loader = torch.utils.data.DataLoader(Dataset(train_x,train_y),batch_size=16,shuffle=True,num_workers=2) val_loader = torch.utils.data.DataLoader(Dataset(val_x,val_y),batch_size=16,shuffle=False,num_workers=2 )<normalization>
rf.fit(X, y) estimator = rf.estimators_[0] export_graphviz(estimator, out_file='tree.dot', feature_names = X.columns, class_names = ['Not Survived','Survived'], rounded = True, proportion = False, precision = 2, filled = True) call(['dot', '-Tpng', 'tree.dot', '-o', 'tree.png', '-Gdpi=600']) plt.figure(figsize =(40, 20)) plt.imshow(plt.imread('tree.png')) plt.axis('off'); plt.show() ;
Titanic - Machine Learning from Disaster
9,251,987
class Model(nn.Module): def __init__(self): super(Model,self ).__init__() self.distilBert = transformers.RobertaModel.from_pretrained('roberta-base') self.l0 = nn.Linear(768,512) self.l1 = nn.Linear(512,256) self.l2 = nn.Linear(256,1) self.d0 = nn.Dropout(0.5) self.d1 = nn.Dropout(0.5) self.d2 = nn.Dropout(0.5) nn.init.normal_(self.l0.weight,std=0.2) nn.init.normal_(self.l1.weight,std=0.2) nn.init.normal_(self.l2.weight,std=0.2) def forward(self,ids,masks): hid = self.distilBert(ids,attention_mask=masks) hid = hid[0][:,0] x = self.d0(hid) x = self.l0(x) x = F.leaky_relu(x) x = self.d1(x) x = self.l1(x) x = F.leaky_relu(x) x = self.d2(x) x = self.l2(x) return x <choose_model_class>
def f_selector(X, y, est, features): X_train, X_valid, y_train, y_valid = train_test_split(X, y, test_size=0.4, random_state=42) rfe = RFE(estimator=est, n_features_to_select=features, verbose=1) rfe.fit(X_train, y_train) print(dict(zip(X.columns, rfe.ranking_))) print(X.columns[rfe.support_]) acc = accuracy_score(y_valid, rfe.predict(X_valid)) print("{0:.1%} accuracy on test set.".format(acc)) X_red = X[X_train.columns[rfe.support_].to_list() ] X_te_red = X_test[X_train.columns[rfe.support_].to_list() ] return X_red, X_te_red
Titanic - Machine Learning from Disaster
9,251,987
model = Model().to('cuda') criterion = nn.BCEWithLogitsLoss(reduction='mean') optimizer = torch.optim.AdamW(model.parameters() ,lr=3e-5 )<compute_test_metric>
X_sel, X_test_sel = f_selector(X, y, rf, 11 )
Titanic - Machine Learning from Disaster
9,251,987
def accuracy_score(outputs,labels): outputs = torch.round(torch.sigmoid(outputs)) correct =(outputs == labels ).sum().float() return correct/labels.size(0 )<import_modules>
pipe = Pipeline([ ('scaler', StandardScaler()), ('reducer', PCA(n_components=2)) , ]) X_sel_red = pipe.fit_transform(X_sel) X_test_sel_red = pipe.transform(X_test_sel )
Titanic - Machine Learning from Disaster
9,251,987
from tqdm import tqdm<train_on_grid>
def prob_reg(X, y): figure = plt.figure(figsize=(20, 40)) h =.02 i = 1 X_train, X_test, y_train, y_test = \ train_test_split(X, y, test_size=.4, random_state=42) x_min, x_max = X_sel_red[:, 0].min() -.5, X_sel_red[:, 0].max() +.5 y_min, y_max = X_sel_red[:, 1].min() -.5, X_sel_red[:, 1].max() +.5 xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h)) cm = plt.cm.RdYlGn cm_bright = ListedColormap([' ax = plt.subplot(5, 2, i) for clf in estimators: ax = plt.subplot(math.ceil(len(estimators)/ 2), 2, i) clf.fit(X_train, y_train) score = clf.score(X_test, y_test) if hasattr(clf, "decision_function"): Z = clf.decision_function(np.c_[xx.ravel() , yy.ravel() ]) else: Z = clf.predict_proba(np.c_[xx.ravel() , yy.ravel() ])[:, 1] Z = Z.reshape(xx.shape) ax.contourf(xx, yy, Z, cmap=cm, alpha=.8) g = ax.scatter(X_train[:, 0], X_train[:, 1], c=y_train, cmap=cm_bright, edgecolors='k') ax.scatter(X_test[:, 0], X_test[:, 1], c=y_test, cmap=cm_bright, edgecolors='k', alpha=0.6) ax.set_xlim(xx.min() , xx.max()) ax.set_ylim(yy.min() , yy.max()) ax.set_title(clf.__class__.__name__) ax.set_xlabel('PCA 1') ax.set_ylabel('PCA 2') plt.legend(handles=g.legend_elements() [0], labels=['Not Survived', 'Survived'], framealpha=0.3, scatterpoints=1) i += 1 plt.tight_layout() plt.show()
Titanic - Machine Learning from Disaster
9,251,987
epochs = 4 for epoch in range(epochs): epoch_loss = 0. model.train() for data in tqdm(train_loader): ids = data['ids'].cuda() masks = data['masks'].cuda() labels = data['out'].cuda() labels = labels.unsqueeze(1) optimizer.zero_grad() outputs = model(ids,masks) loss = criterion(outputs,labels) loss.backward() optimizer.step() epoch_loss += loss.item() print(f"Epoch{epoch+1} loss : {epoch_loss/len(train_loader)}") print(f"Epoch{epoch+1} accuracy : ",accuracy_score(outputs,labels ).item()) val_loss = 0. model.eval() for data in tqdm(val_loader): ids = data['ids'].cuda() masks = data['masks'].cuda() labels = data['out'].cuda() labels = labels.unsqueeze(1) outputs = model(ids,masks) loss = criterion(outputs,labels) val_loss += loss.item() print(f"Epoch{epoch+1} val loss : {val_loss/len(val_loader)}") print(f"Epoch{epoch+1} val accuracy : ",accuracy_score(outputs,labels ).item()) <create_dataframe>
dec_regs(X_sel_red, y, estimators )
Titanic - Machine Learning from Disaster
9,251,987
test_loader = torch.utils.data.DataLoader(Dataset(test_df['text'],y=None),batch_size=16,shuffle=False,num_workers=2 )<find_best_params>
prob_reg(X_sel_red, y )
Titanic - Machine Learning from Disaster
9,251,987
preds = [] for data in test_loader: ids = data['ids'].cuda() masks = data['masks'].cuda() model.eval() outputs = model(ids,masks) preds += outputs.cpu().detach().numpy().tolist() <prepare_output>
pca_models = model_check(X_sel_red, y, estimators, cv) display(pca_models.style.background_gradient(cmap='summer_r'))
Titanic - Machine Learning from Disaster
9,251,987
pred = np.round(1/(1 + np.exp(-np.array(preds))))<prepare_output>
rand_model_full_data = rf.fit(X, y) print(accuracy_score(y, rand_model_full_data.predict(X))) y_pred = rand_model_full_data.predict(X_test )
Titanic - Machine Learning from Disaster
9,251,987
<load_from_csv><EOS>
test_df = pd.read_csv('/kaggle/input/titanic/test.csv') submission_df = pd.DataFrame(columns=['PassengerId', 'Survived']) submission_df['PassengerId'] = test_df['PassengerId'] submission_df['Survived'] = y_pred submission_df.to_csv('submission.csv', header=True, index=False) submission_df.head(10 )
Titanic - Machine Learning from Disaster
1,373,360
<SOS> metric: categorizationaccuracy Kaggle data source: titanic-machine-learning-from-disaster<prepare_output>
class KaggleMember() : kaggle_member_count=0 def __init__(self, name, surname, level=None): self.name=name.capitalize() self.surname=surname.upper() self._set_level(level) KaggleMember.kaggle_member_count+=1 self.kaggle_id=KaggleMember.kaggle_member_count def display_number_of_member(self): print("There are {} members in Kaggle community".format(KaggleMember.kaggle_member_count)) def display_member_info(self): print("Kaggle Member Full Name:{} {}".format(self.name, self.surname)) print("Level:{:15}".format(self.level)) print("Kaggle ID:",self.kaggle_id) def _set_level(self, level): if level is None: self.level='Novice' else: self.level=level.title() myclass_name='Kaggle Member'
Titanic - Machine Learning from Disaster
1,373,360
sub['target'] = pred<save_to_csv>
warnings.filterwarnings('ignore') print("Warnings were ignored" )
Titanic - Machine Learning from Disaster
1,373,360
sub.to_csv('submission.csv',index=False )<set_options>
class Information() : def __init__(self): print("Information object created") def _get_missing_values(self,data): missing_values = data.isnull().sum() missing_values.sort_values(ascending=False, inplace=True) return missing_values def info(self,data): feature_dtypes=data.dtypes self.missing_values=self._get_missing_values(data) print("=" * 50) print("{:16} {:16} {:25} {:16}".format("Feature Name".upper() , "Data Format".upper() , " "Samples".upper())) for feature_name, dtype, missing_value in zip(self.missing_values.index.values, feature_dtypes[self.missing_values.index.values], self.missing_values.values): print("{:18} {:19} {:19} ".format(feature_name, str(dtype), str(missing_value)) , end="") for v in data[feature_name].values[:10]: print(v, end=",") print() print("="*50)
Titanic - Machine Learning from Disaster
1,373,360
SEED = 42 torch.manual_seed(SEED) torch.backends.cudnn.deterministic = True torch.backends.cudnn.benchmark = False<load_from_csv>
class Preprocess() : def __init__(self): print("Preprocess object created") def fillna(self, data, fill_strategies): for column, strategy in fill_strategies.items() : if strategy == 'None': data[column] = data[column].fillna('None') elif strategy == 'Zero': data[column] = data[column].fillna(0) elif strategy == 'Mode': data[column] = data[column].fillna(data[column].mode() [0]) elif strategy == 'Mean': data[column] = data[column].fillna(data[column].mean()) elif strategy == 'Median': data[column] = data[column].fillna(data[column].median()) else: print("{}: There is no such thing as preprocess strategy".format(strategy)) return data def drop(self, data, drop_strategies): for column, strategy in drop_strategies.items() : data=data.drop(labels=[column], axis=strategy) return data def feature_engineering(self, data, engineering_strategies=1): if engineering_strategies==1: return self._feature_engineering1(data) return data def _feature_engineering1(self,data): data=self._base_feature_engineering(data) data['FareBin'] = pd.qcut(data['Fare'], 4) data['AgeBin'] = pd.cut(data['Age'].astype(int), 5) drop_strategy = {'Age': 1, 'Name': 1, 'Fare': 1} data = self.drop(data, drop_strategy) return data def _base_feature_engineering(self,data): data['FamilySize'] = data['SibSp'] + data['Parch'] + 1 data['IsAlone'] = 1 data.loc[(data['FamilySize'] > 1), 'IsAlone'] = 0 data['Title'] = data['Name'].str.split(", ", expand=True)[1].str.split('.', expand=True)[0] min_lengtht = 10 title_names =(data['Title'].value_counts() < min_lengtht) data['Title'] = data['Title'].apply(lambda x: 'Misc' if title_names.loc[x] == True else x) return data def _label_encoder(self,data): labelEncoder=LabelEncoder() for column in data.columns.values: if 'int64'==data[column].dtype or 'float64'==data[column].dtype or 'int64'==data[column].dtype: continue labelEncoder.fit(data[column]) data[column]=labelEncoder.transform(data[column]) return data def _get_dummies(self, data, prefered_columns=None): if prefered_columns is None: columns=data.columns.values non_dummies=None else: non_dummies=[col for col in data.columns.values if col not in prefered_columns ] columns=prefered_columns dummies_data=[pd.get_dummies(data[col],prefix=col)for col in columns] if non_dummies is not None: for non_dummy in non_dummies: dummies_data.append(data[non_dummy]) return pd.concat(dummies_data, axis=1 )
Titanic - Machine Learning from Disaster
1,373,360
train_data = pd.read_csv(".. /input/nlp-getting-started/train.csv") train_data.info() train_data.sample(10 )<load_from_csv>
class PreprocessStrategy() : def __init__(self): self.data=None self._preprocessor=Preprocess() def strategy(self, data, strategy_type="strategy1"): self.data=data if strategy_type=='strategy1': self._strategy1() elif strategy_type=='strategy2': self._strategy2() return self.data def _base_strategy(self): drop_strategy = {'PassengerId': 1, 'Cabin': 1, 'Ticket': 1} self.data = self._preprocessor.drop(self.data, drop_strategy) fill_strategy = {'Age': 'Median', 'Fare': 'Median', 'Embarked': 'Mode'} self.data = self._preprocessor.fillna(self.data, fill_strategy) self.data = self._preprocessor.feature_engineering(self.data, 1) self.data = self._preprocessor._label_encoder(self.data) def _strategy1(self): self._base_strategy() self.data=self._preprocessor._get_dummies(self.data, prefered_columns=['Pclass', 'Sex', 'Parch', 'Embarked', 'Title', 'IsAlone']) def _strategy2(self): self._base_strategy() self.data=self._preprocessor._get_dummies(self.data, prefered_columns=None )
Titanic - Machine Learning from Disaster
1,373,360
test_data = pd.read_csv(".. /input/nlp-getting-started/test.csv") test_data.info() test_data.sample(10 )<train_model>
class GridSearchHelper() : def __init__(self): print("GridSearchHelper Created") self.gridSearchCV=None self.clf_and_params=list() self._initialize_clf_and_params() def _initialize_clf_and_params(self): clf= KNeighborsClassifier() params={'n_neighbors':[5,7,9,11,13,15], 'leaf_size':[1,2,3,5], 'weights':['uniform', 'distance'] } self.clf_and_params.append(( clf, params)) clf=LogisticRegression() params={'penalty':['l1', 'l2'], 'C':np.logspace(0, 4, 10) } self.clf_and_params.append(( clf, params)) clf = SVC() params = [ {'C': [1, 10, 100, 1000], 'kernel': ['linear']}, {'C': [1, 10, 100, 1000], 'gamma': [0.001, 0.0001], 'kernel': ['rbf']}] self.clf_and_params.append(( clf, params)) clf=DecisionTreeClassifier() params={'max_features': ['auto', 'sqrt', 'log2'], 'min_samples_split': [2,3,4,5,6,7,8,9,10,11,12,13,14,15], 'min_samples_leaf':[1], 'random_state':[123]} self.clf_and_params.append(( clf,params)) clf = RandomForestClassifier() params = {'n_estimators': [4, 6, 9], 'max_features': ['log2', 'sqrt','auto'], 'criterion': ['entropy', 'gini'], 'max_depth': [2, 3, 5, 10], 'min_samples_split': [2, 3, 5], 'min_samples_leaf': [1,5,8] } self.clf_and_params.append(( clf, params)) def fit_predict_save(self, X_train, X_test, y_train, submission_id, strategy_type): self.X_train=X_train self.X_test=X_test self.y_train=y_train self.submission_id=submission_id self.strategy_type=strategy_type clf_and_params = self.get_clf_and_params() models=[] self.results={} for clf, params in clf_and_params: self.current_clf_name = clf.__class__.__name__ grid_search_clf = GridSearchCV(clf, params, cv=5) grid_search_clf.fit(self.X_train, self.y_train) self.Y_pred = grid_search_clf.predict(self.X_test) clf_train_acc = round(grid_search_clf.score(self.X_train, self.y_train)* 100, 2) print(self.current_clf_name, " trained and used for prediction on test data...") self.results[self.current_clf_name]=clf_train_acc models.append(clf) self.save_result() print() def show_result(self): for clf_name, train_acc in self.results.items() : print("{} train accuracy is {:.3f}".format(clf_name, train_acc)) def save_result(self): Submission = pd.DataFrame({'PassengerId': self.submission_id, 'Survived': self.Y_pred}) file_name="{}_{}.csv".format(self.strategy_type,self.current_clf_name.lower()) Submission.to_csv(file_name, index=False) print("Submission saved file name: ",file_name) def get_clf_and_params(self): return self.clf_and_params def add(self,clf, params): self.clf_and_params.append(( clf, params))
Titanic - Machine Learning from Disaster
1,373,360
print('Training Set Shape = {}'.format(train_data.shape)) print('Test Set Shape = {}'.format(test_data.shape))<count_unique_values>
class Visualizer: def __init__(self): print("Visualizer object created!") def RandianViz(self, X, y, number_of_features): if number_of_features is None: features=X.columns.values else: features=X.columns.values[:number_of_features] fig, ax=plt.subplots(1, figsize=(15,12)) radViz=RadViz(classes=['survived', 'not survived'], features=features) radViz.fit(X, y) radViz.transform(X) radViz.poof()
Titanic - Machine Learning from Disaster
1,373,360
mislabeled_df = train_data.groupby(['text'] ).nunique().sort_values(by='target', ascending=False) mislabeled_df = mislabeled_df[mislabeled_df['target'] > 1]['target'] mislabeled_list = mislabeled_df.index.tolist() mislabeled_list<feature_engineering>
class ObjectOrientedTitanic() : def __init__(self, train, test): print("ObjectOrientedTitanic object created") self.testPassengerID=test['PassengerId'] self.number_of_train=train.shape[0] self.y_train=train['Survived'] self.train=train.drop('Survived', axis=1) self.test=test self.all_data=self._get_all_data() self._info=Information() self.preprocessStrategy = PreprocessStrategy() self.visualizer=Visualizer() self.gridSearchHelper = GridSearchHelper() def _get_all_data(self): return pd.concat([self.train, self.test]) def information(self): self._info.info(self.all_data) def preprocessing(self, strategy_type): self.strategy_type=strategy_type self.all_data = self.preprocessStrategy.strategy(self._get_all_data() , strategy_type) def visualize(self, visualizer_type, number_of_features=None): self._get_train_and_test() if visualizer_type=="RadViz": self.visualizer.RandianViz(X=self.X_train, y=self.y_train, number_of_features=number_of_features) def machine_learning(self): self._get_train_and_test() self.gridSearchHelper.fit_predict_save(self.X_train, self.X_test, self.y_train, self.testPassengerID, self.strategy_type) def show_result(self): self.gridSearchHelper.show_result() def _get_train_and_test(self): self.X_train=self.all_data[:self.number_of_train] self.X_test=self.all_data[self.number_of_train:]
Titanic - Machine Learning from Disaster
1,373,360
train_data['target_relabeled'] = train_data['target'].copy() train_data.loc[train_data['text'] == 'like for the music video I want some real action shit like burning buildings and police chases not some weak ben winston shit', 'target_relabeled'] = 0 train_data.loc[train_data['text'] == 'Hellfire is surrounded by desires so be careful and don‰Ûªt let your desires control you! train_data.loc[train_data['text'] == 'To fight bioterrorism sir.', 'target_relabeled'] = 0 train_data.loc[train_data['text'] == '.POTUS train_data.loc[train_data['text'] == 'CLEARED:incident with injury:I-495 inner loop Exit 31 - MD 97/Georgia Ave Silver Spring', 'target_relabeled'] = 1 train_data.loc[train_data['text'] == ' train_data.loc[train_data['text'] == 'In train_data.loc[train_data['text'] == 'Who is bringing the tornadoes and floods.Who is bringing the climate change.God is after America He is plaguing her train_data.loc[train_data['text'] == 'RT NotExplained: The only known image of infamous hijacker D.B.Cooper.http://t.co/JlzK2HdeTG', 'target_relabeled'] = 1 train_data.loc[train_data['text'] == "Mmmmmm I'm burning.... I'm burning buildings I'm building.... Oooooohhhh oooh ooh...", 'target_relabeled'] = 0 train_data.loc[train_data['text'] == "wowo--=== 12000 Nigerian refugees repatriated from Cameroon", 'target_relabeled'] = 0 train_data.loc[train_data['text'] == "He came to a land which was engulfed in tribal war and turned it into a land of peace i.e.Madinah. train_data.loc[train_data['text'] == "Hellfire! We don‰Ûªt even want to think about it or mention it so let‰Ûªs not do anything that leads to it train_data.loc[train_data['text'] == "The Prophet(peace be upon him)said 'Save yourself from Hellfire even if it is by giving half a date in charity.'", 'target_relabeled'] = 0 train_data.loc[train_data['text'] == "Caution: breathing may be hazardous to your health.", 'target_relabeled'] = 1 train_data.loc[train_data['text'] == "I Pledge Allegiance To The P.O.P.E.And The Burning Buildings of Epic City.??????", 'target_relabeled'] = 0 train_data.loc[train_data['text'] == " train_data.loc[train_data['text'] == "that horrible sinking feeling when you‰Ûªve been at home on your phone for a while and you realise its been on 3G this whole time", 'target_relabeled'] = 0<categorify>
train = pd.read_csv(".. /input/train.csv") test = pd.read_csv(".. /input/test.csv") objectOrientedTitanic=ObjectOrientedTitanic(train, test)
Titanic - Machine Learning from Disaster
1,373,360
<split><EOS>
objectOrientedTitanic.machine_learning()
Titanic - Machine Learning from Disaster
2,725,427
<SOS> metric: categorizationaccuracy Kaggle data source: titanic-machine-learning-from-disaster<define_variables>
sns.set(style="darkgrid") warnings.filterwarnings('ignore') SEED = 42
Titanic - Machine Learning from Disaster
2,725,427
TEXT = data.Field(tokenize = 'spacy', batch_first=True, include_lengths = True) LABEL = data.LabelField(dtype = torch.float, batch_first=True )<split>
def concat_df(train_data, test_data): return pd.concat([train_data, test_data], sort=True ).reset_index(drop=True) def divide_df(all_data): return all_data.loc[:890], all_data.loc[891:].drop(['Survived'], axis=1) df_train = pd.read_csv('.. /input/train.csv') df_test = pd.read_csv('.. /input/test.csv') df_all = concat_df(df_train, df_test) df_train.name = 'Training Set' df_test.name = 'Test Set' df_all.name = 'All Set' dfs = [df_train, df_test] print('Number of Training Examples = {}'.format(df_train.shape[0])) print('Number of Test Examples = {} '.format(df_test.shape[0])) print('Training X Shape = {}'.format(df_train.shape)) print('Training y Shape = {} '.format(df_train['Survived'].shape[0])) print('Test X Shape = {}'.format(df_test.shape)) print('Test y Shape = {} '.format(df_test.shape[0])) print(df_train.columns) print(df_test.columns )
Titanic - Machine Learning from Disaster
2,725,427
class DataFrameDataset(data.Dataset): def __init__(self, df, fields, is_test=False, **kwargs): examples = [] for i, row in df.iterrows() : label = row.target_relabeled if not is_test else None text = row.text examples.append(data.Example.fromlist([text, label], fields)) super().__init__(examples, fields, **kwargs) @staticmethod def sort_key(ex): return len(ex.text) @classmethod def splits(cls, fields, train_df, val_df=None, test_df=None, **kwargs): train_data, val_data, test_data =(None, None, None) data_field = fields if train_df is not None: train_data = cls(train_df.copy() , data_field, **kwargs) if val_df is not None: val_data = cls(val_df.copy() , data_field, **kwargs) if test_df is not None: test_data = cls(test_df.copy() , data_field, False, **kwargs) return tuple(d for d in(train_data, val_data, test_data)if d is not None )<create_dataframe>
def display_missing(df): for col in df.columns.tolist() : print('{} column missing values: {}'.format(col, df[col].isnull().sum())) print(' ') for df in dfs: print('{}'.format(df.name)) display_missing(df )
Titanic - Machine Learning from Disaster
2,725,427
fields = [('text',TEXT),('label',LABEL)] train_ds, val_ds = DataFrameDataset.splits(fields, train_df=train_df, val_df=valid_df )<load_pretrained>
age_by_pclass_sex = df_all.groupby(['Sex', 'Pclass'] ).median() ['Age'] for pclass in range(1, 4): for sex in ['female', 'male']: print('Median age of Pclass {} {}s: {}'.format(pclass, sex, age_by_pclass_sex[sex][pclass])) print('Median age of all passengers: {}'.format(df_all['Age'].median())) df_all['Age'] = df_all.groupby(['Sex', 'Pclass'])['Age'].apply(lambda x: x.fillna(x.median()))
Titanic - Machine Learning from Disaster
2,725,427
vectors = Vectors(name='.. /input/fasttext-crawl-300d-2m/crawl-300d-2M.vec', cache='./') MAX_VOCAB_SIZE = 100000 TEXT.build_vocab(train_ds, max_size = MAX_VOCAB_SIZE, vectors = vectors, unk_init = torch.Tensor.zero_) LABEL.build_vocab(train_ds )<count_unique_values>
df_all[df_all['Embarked'].isnull() ]
Titanic - Machine Learning from Disaster
2,725,427
print("Size of TEXT vocabulary:",len(TEXT.vocab)) print("Size of LABEL vocabulary:",len(LABEL.vocab)) print(TEXT.vocab.freqs.most_common(10)) <split>
df_all['Embarked'] = df_all['Embarked'].fillna('S' )
Titanic - Machine Learning from Disaster
2,725,427
BATCH_SIZE = 64 device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') train_iterator, valid_iterator = data.BucketIterator.splits( (train_ds, val_ds), batch_size = BATCH_SIZE, sort_within_batch = True, device = device )<init_hyperparams>
df_all[df_all['Fare'].isnull() ]
Titanic - Machine Learning from Disaster
2,725,427
num_epochs = 25 learning_rate = 0.001 INPUT_DIM = len(TEXT.vocab) EMBEDDING_DIM = 300 HIDDEN_DIM = 256 OUTPUT_DIM = 1 N_LAYERS = 2 BIDIRECTIONAL = True DROPOUT = 0.2 PAD_IDX = TEXT.vocab.stoi[TEXT.pad_token]<choose_model_class>
med_fare = df_all.groupby(['Pclass', 'Parch', 'SibSp'] ).Fare.median() [3][0][0] df_all['Fare'] = df_all['Fare'].fillna(med_fare )
Titanic - Machine Learning from Disaster
2,725,427
class LSTM_net(nn.Module): def __init__(self, vocab_size, embedding_dim, hidden_dim, output_dim, n_layers, bidirectional, dropout, pad_idx): super().__init__() self.embedding = nn.Embedding(vocab_size, embedding_dim, padding_idx = pad_idx) self.rnn = nn.LSTM(embedding_dim, hidden_dim, num_layers=n_layers, bidirectional=bidirectional, dropout=dropout, batch_first=True) self.fc1 = nn.Linear(hidden_dim * 2, hidden_dim) self.fc2 = nn.Linear(hidden_dim, 1) self.dropout = nn.Dropout(dropout) def forward(self, text, text_lengths): embedded = self.embedding(text) packed_embedded = nn.utils.rnn.pack_padded_sequence(embedded, text_lengths, batch_first=True) packed_output,(hidden, cell)= self.rnn(packed_embedded) hidden = self.dropout(torch.cat(( hidden[-2,:,:], hidden[-1,:,:]), dim = 1)) output = self.fc1(hidden) output = self.dropout(self.fc2(output)) return output<choose_model_class>
idx = df_all[df_all['Deck'] == 'T'].index df_all.loc[idx, 'Deck'] = 'A'
Titanic - Machine Learning from Disaster
2,725,427
model = LSTM_net(INPUT_DIM, EMBEDDING_DIM, HIDDEN_DIM, OUTPUT_DIM, N_LAYERS, BIDIRECTIONAL, DROPOUT, PAD_IDX )<find_best_params>
df_all_decks_survived = df_all.groupby(['Deck', 'Survived'] ).count().drop(columns=['Sex', 'Age', 'SibSp', 'Parch', 'Fare', 'Embarked', 'Pclass', 'Cabin', 'PassengerId', 'Ticket'] ).rename(columns={'Name':'Count'} ).transpose() def get_survived_dist(df): surv_counts = {'A':{}, 'B':{}, 'C':{}, 'D':{}, 'E':{}, 'F':{}, 'G':{}, 'M':{}} decks = df.columns.levels[0] for deck in decks: for survive in range(0, 2): surv_counts[deck][survive] = df[deck][survive][0] df_surv = pd.DataFrame(surv_counts) surv_percentages = {} for col in df_surv.columns: surv_percentages[col] = [(count / df_surv[col].sum())* 100 for count in df_surv[col]] return surv_counts, surv_percentages def display_surv_dist(percentages): df_survived_percentages = pd.DataFrame(percentages ).transpose() deck_names =('A', 'B', 'C', 'D', 'E', 'F', 'G', 'M') bar_count = np.arange(len(deck_names)) bar_width = 0.85 not_survived = df_survived_percentages[0] survived = df_survived_percentages[1] plt.figure(figsize=(20, 10)) plt.bar(bar_count, not_survived, color=' plt.bar(bar_count, survived, bottom=not_survived, color=' plt.xlabel('Deck', size=15, labelpad=20) plt.ylabel('Survival Percentage', size=15, labelpad=20) plt.xticks(bar_count, deck_names) plt.tick_params(axis='x', labelsize=15) plt.tick_params(axis='y', labelsize=15) plt.legend(loc='upper left', bbox_to_anchor=(1, 1), prop={'size': 15}) plt.title('Survival Percentage in Decks', size=18, y=1.05) plt.show() all_surv_count, all_surv_per = get_survived_dist(df_all_decks_survived) display_surv_dist(all_surv_per )
Titanic - Machine Learning from Disaster
2,725,427
print(model) def count_parameters(model): return sum(p.numel() for p in model.parameters() if p.requires_grad) print(f'The model has {count_parameters(model):,} trainable parameters' )<feature_engineering>
df_all['Deck'] = df_all['Deck'].replace(['A', 'B', 'C'], 'ABC') df_all['Deck'] = df_all['Deck'].replace(['D', 'E'], 'DE') df_all['Deck'] = df_all['Deck'].replace(['F', 'G'], 'FG') df_all['Deck'].value_counts()
Titanic - Machine Learning from Disaster
2,725,427
pretrained_embeddings = TEXT.vocab.vectors model.embedding.weight.data.copy_(pretrained_embeddings) model.embedding.weight.data[PAD_IDX] = torch.zeros(EMBEDDING_DIM )<compute_test_metric>
df_all.drop(['Cabin'], inplace=True, axis=1) df_train, df_test = divide_df(df_all) dfs = [df_train, df_test] for df in dfs: display_missing(df )
Titanic - Machine Learning from Disaster
2,725,427
def binary_accuracy(preds, y): rounded_preds = torch.round(torch.sigmoid(preds)) correct =(rounded_preds == y ).float() acc = correct.sum() / len(correct) return acc<train_on_grid>
corr = df_train_corr_nd['Correlation Coefficient'] > 0.1 df_train_corr_nd[corr]
Titanic - Machine Learning from Disaster
2,725,427
def train(model, iterator, optimizer, criterion): epoch_loss = 0 epoch_acc = 0 model.train() for batch in iterator: text, text_lengths = batch.text optimizer.zero_grad() predictions = model(text, text_lengths ).squeeze(1) loss = criterion(predictions, batch.label) acc = binary_accuracy(predictions, batch.label) loss.backward() optimizer.step() epoch_loss += loss.item() epoch_acc += acc.item() return epoch_loss / len(iterator), epoch_acc / len(iterator )<compute_train_metric>
corr = df_test_corr_nd['Correlation Coefficient'] > 0.1 df_test_corr_nd[corr]
Titanic - Machine Learning from Disaster
2,725,427
def evaluate(model, iterator, criterion): epoch_loss = 0 epoch_acc = 0 model.eval() with torch.no_grad() : for batch in iterator: text, text_lengths = batch.text predictions = model(text, text_lengths ).squeeze(1) loss = criterion(predictions, batch.label) acc = binary_accuracy(predictions, batch.label) epoch_loss += loss.item() epoch_acc += acc.item() return epoch_loss / len(iterator), epoch_acc / len(iterator )<choose_model_class>
df_all = concat_df(df_train, df_test) df_all.head()
Titanic - Machine Learning from Disaster
2,725,427
t = time.time() best_valid_loss = float('inf') model.to(device) criterion = nn.BCEWithLogitsLoss() optimizer = torch.optim.Adam(model.parameters() , lr=learning_rate) for epoch in range(num_epochs): train_loss, train_acc = train(model, train_iterator, optimizer, criterion) valid_loss, valid_acc = evaluate(model, valid_iterator, criterion) print(f'\tTrain Loss: {train_loss:.4f} | Train Acc: {train_acc*100:.4f}%') print(f'\t Val.Loss: {valid_loss:.4f} | Val.Acc: {valid_acc*100:.4f}%') if valid_loss < best_valid_loss: best_valid_loss = valid_loss torch.save(model.state_dict() , 'best_model.pt') print(f'time:{time.time() -t:.3f}' )<predict_on_test>
df_all['Fare'] = pd.qcut(df_all['Fare'], 13 )
Titanic - Machine Learning from Disaster
2,725,427
nlp = spacy.load('en') def predict(model, sentence): tokenized = [tok.text for tok in nlp.tokenizer(sentence)] indexed = [TEXT.vocab.stoi[t] for t in tokenized] length = [len(indexed)] tensor = torch.LongTensor(indexed ).to(device) tensor = tensor.unsqueeze(1 ).T length_tensor = torch.LongTensor(length) prediction = model(tensor, length_tensor ).squeeze(1) rounded_preds = torch.round(torch.sigmoid(prediction)) predict_class = rounded_preds.tolist() [0] return predict_class<predict_on_test>
df_all['Age'] = pd.qcut(df_all['Age'], 10 )
Titanic - Machine Learning from Disaster
2,725,427
PATH = ".. /working/best_model.pt" model.load_state_dict(torch.load(PATH)) predicts = [] for i in range(len(test_data.text)) : predict_class = predict(model, test_data.text[i]) predicts.append(int(predict_class))<load_from_csv>
df_all['Ticket_Frequency'] = df_all.groupby('Ticket')['Ticket'].transform('count' )
Titanic - Machine Learning from Disaster
2,725,427
submission = pd.read_csv(".. /input/nlp-getting-started/sample_submission.csv") submission['target'] = predicts submission<save_to_csv>
df_all['Title'] = df_all['Name'].str.split(', ', expand=True)[1].str.split('.', expand=True)[0] df_all['Is_Married'] = 0 df_all['Is_Married'].loc[df_all['Title'] == 'Mrs'] = 1
Titanic - Machine Learning from Disaster
2,725,427
submission.to_csv('submission.csv',index=False )<merge>
def extract_surname(data): families = [] for i in range(len(data)) : name = data.iloc[i] if '(' in name: name_no_bracket = name.split('(')[0] else: name_no_bracket = name family = name_no_bracket.split(',')[0] title = name_no_bracket.split(',')[1].strip().split(' ')[0] for c in string.punctuation: family = family.replace(c, '' ).strip() families.append(family) return families df_all['Family'] = extract_surname(df_all['Name']) df_train = df_all.loc[:890] df_test = df_all.loc[891:] dfs = [df_train, df_test]
Titanic - Machine Learning from Disaster
2,725,427
gt_df = pd.read_csv(".. /input/disasters-on-social-media/socialmedia-disaster-tweets-DFE.csv", encoding='latin_1') gt_df = gt_df[['choose_one', 'text']] gt_df['target'] =(gt_df['choose_one']=='Relevant' ).astype(int) gt_df['id'] = gt_df.index merged_df = pd.merge(test_data, gt_df, on='id') merged_df<prepare_x_and_y>
mean_survival_rate = np.mean(df_train['Survived']) train_family_survival_rate = [] train_family_survival_rate_NA = [] test_family_survival_rate = [] test_family_survival_rate_NA = [] for i in range(len(df_train)) : if df_train['Family'][i] in family_rates: train_family_survival_rate.append(family_rates[df_train['Family'][i]]) train_family_survival_rate_NA.append(1) else: train_family_survival_rate.append(mean_survival_rate) train_family_survival_rate_NA.append(0) for i in range(len(df_test)) : if df_test['Family'].iloc[i] in family_rates: test_family_survival_rate.append(family_rates[df_test['Family'].iloc[i]]) test_family_survival_rate_NA.append(1) else: test_family_survival_rate.append(mean_survival_rate) test_family_survival_rate_NA.append(0) df_train['Family_Survival_Rate'] = train_family_survival_rate df_train['Family_Survival_Rate_NA'] = train_family_survival_rate_NA df_test['Family_Survival_Rate'] = test_family_survival_rate df_test['Family_Survival_Rate_NA'] = test_family_survival_rate_NA train_ticket_survival_rate = [] train_ticket_survival_rate_NA = [] test_ticket_survival_rate = [] test_ticket_survival_rate_NA = [] for i in range(len(df_train)) : if df_train['Ticket'][i] in ticket_rates: train_ticket_survival_rate.append(ticket_rates[df_train['Ticket'][i]]) train_ticket_survival_rate_NA.append(1) else: train_ticket_survival_rate.append(mean_survival_rate) train_ticket_survival_rate_NA.append(0) for i in range(len(df_test)) : if df_test['Ticket'].iloc[i] in ticket_rates: test_ticket_survival_rate.append(ticket_rates[df_test['Ticket'].iloc[i]]) test_ticket_survival_rate_NA.append(1) else: test_ticket_survival_rate.append(mean_survival_rate) test_ticket_survival_rate_NA.append(0) df_train['Ticket_Survival_Rate'] = train_ticket_survival_rate df_train['Ticket_Survival_Rate_NA'] = train_ticket_survival_rate_NA df_test['Ticket_Survival_Rate'] = test_ticket_survival_rate df_test['Ticket_Survival_Rate_NA'] = test_ticket_survival_rate_NA
Titanic - Machine Learning from Disaster
2,725,427
target_df = merged_df[['id', 'target']] target_df<save_to_csv>
for df in [df_train, df_test]: df['Survival_Rate'] =(df['Ticket_Survival_Rate'] + df['Family_Survival_Rate'])/ 2 df['Survival_Rate_NA'] =(df['Ticket_Survival_Rate_NA'] + df['Family_Survival_Rate_NA'])/ 2
Titanic - Machine Learning from Disaster
2,725,427
target_df.to_csv('perfect_submission.csv', index=False )<compute_test_metric>
non_numeric_features = ['Embarked', 'Sex', 'Deck', 'Title', 'Family_Size_Grouped', 'Age', 'Fare'] for df in dfs: for feature in non_numeric_features: df[feature] = LabelEncoder().fit_transform(df[feature] )
Titanic - Machine Learning from Disaster
2,725,427
target_df["predict"] = list(submission.target) print('\t\tCLASSIFICATIION METRICS ') print(metrics.classification_report(target_df.target, target_df.predict))<set_options>
cat_features = ['Pclass', 'Sex', 'Deck', 'Embarked', 'Title', 'Family_Size_Grouped'] encoded_features = [] for df in dfs: for feature in cat_features: encoded_feat = OneHotEncoder().fit_transform(df[feature].values.reshape(-1, 1)).toarray() n = df[feature].nunique() cols = ['{}_{}'.format(feature, n)for n in range(1, n + 1)] encoded_df = pd.DataFrame(encoded_feat, columns=cols) encoded_df.index = df.index encoded_features.append(encoded_df) df_train = pd.concat([df_train, *encoded_features[:6]], axis=1) df_test = pd.concat([df_test, *encoded_features[6:]], axis=1 )
Titanic - Machine Learning from Disaster
2,725,427
plt.style.use('ggplot') warnings.filterwarnings('ignore') <define_variables>
df_all = concat_df(df_train, df_test) drop_cols = ['Deck', 'Embarked', 'Family', 'Family_Size', 'Family_Size_Grouped', 'Survived', 'Name', 'Parch', 'PassengerId', 'Pclass', 'Sex', 'SibSp', 'Ticket', 'Title', 'Ticket_Survival_Rate', 'Family_Survival_Rate', 'Ticket_Survival_Rate_NA', 'Family_Survival_Rate_NA'] df_all.drop(columns=drop_cols, inplace=True) df_all.head()
Titanic - Machine Learning from Disaster
2,725,427
def seed_everything(seed): os.environ['PYTHONHASHSEED']=str(seed) tf.random.set_seed(seed) np.random.seed(seed) random.seed(seed) seed_everything(34 )<load_from_csv>
X_train = StandardScaler().fit_transform(df_train.drop(columns=drop_cols)) y_train = df_train['Survived'].values X_test = StandardScaler().fit_transform(df_test.drop(columns=drop_cols)) print('X_train shape: {}'.format(X_train.shape)) print('y_train shape: {}'.format(y_train.shape)) print('X_test shape: {}'.format(X_test.shape))
Titanic - Machine Learning from Disaster
2,725,427
train = pd.read_csv('.. /input/nlp-getting-started/train.csv') test = pd.read_csv('.. /input/nlp-getting-started/test.csv') train.head()<categorify>
single_best_model = RandomForestClassifier(criterion='gini', n_estimators=1100, max_depth=5, min_samples_split=4, min_samples_leaf=5, max_features='auto', oob_score=True, random_state=SEED, n_jobs=-1, verbose=1) leaderboard_model = RandomForestClassifier(criterion='gini', n_estimators=1750, max_depth=7, min_samples_split=6, min_samples_leaf=6, max_features='auto', oob_score=True, random_state=SEED, n_jobs=-1, verbose=1 )
Titanic - Machine Learning from Disaster
2,725,427
test_id = test['id'] columns = {'id', 'location'} train = train.drop(columns = columns) test = test.drop(columns = columns) train['keyword'] = train['keyword'].fillna('unknown') test['keyword'] = test['keyword'].fillna('unknown') train['text'] = train['text'] + ' ' + train['keyword'] test['text'] = test['text'] + ' ' + test['keyword'] columns = {'keyword'} train = train.drop(columns = columns) test = test.drop(columns = columns) total = train.append(test )<feature_engineering>
N = 5 oob = 0 probs = pd.DataFrame(np.zeros(( len(X_test), N * 2)) , columns=['Fold_{}_Prob_{}'.format(i, j)for i in range(1, N + 1)for j in range(2)]) importances = pd.DataFrame(np.zeros(( X_train.shape[1], N)) , columns=['Fold_{}'.format(i)for i in range(1, N + 1)], index=df_all.columns) fprs, tprs, scores = [], [], [] skf = StratifiedKFold(n_splits=N, random_state=N, shuffle=True) for fold,(trn_idx, val_idx)in enumerate(skf.split(X_train, y_train), 1): print('Fold {} '.format(fold)) leaderboard_model.fit(X_train[trn_idx], y_train[trn_idx]) trn_fpr, trn_tpr, trn_thresholds = roc_curve(y_train[trn_idx], leaderboard_model.predict_proba(X_train[trn_idx])[:, 1]) trn_auc_score = auc(trn_fpr, trn_tpr) val_fpr, val_tpr, val_thresholds = roc_curve(y_train[val_idx], leaderboard_model.predict_proba(X_train[val_idx])[:, 1]) val_auc_score = auc(val_fpr, val_tpr) scores.append(( trn_auc_score, val_auc_score)) fprs.append(val_fpr) tprs.append(val_tpr) probs.loc[:, 'Fold_{}_Prob_0'.format(fold)] = leaderboard_model.predict_proba(X_test)[:, 0] probs.loc[:, 'Fold_{}_Prob_1'.format(fold)] = leaderboard_model.predict_proba(X_test)[:, 1] importances.iloc[:, fold - 1] = leaderboard_model.feature_importances_ oob += leaderboard_model.oob_score_ / N print('Fold {} OOB Score: {} '.format(fold, leaderboard_model.oob_score_)) print('Average OOB Score: {}'.format(oob))
Titanic - Machine Learning from Disaster
2,725,427
total['unique word count'] = total['text'].apply(lambda x: len(set(x.split()))) total['stopword count'] = total['text'].apply(lambda x: len([i for i in x.lower().split() if i in wordcloud.STOPWORDS])) total['stopword ratio'] = total['stopword count'] / total['word count'] total['punctuation count'] = total['text'].apply(lambda x: len([i for i in str(x)if i in string.punctuation])) train = total[:len(train)] disaster = train['target'] == 1 fig, axes = plt.subplots(4, figsize=(20, 30)) graph1 = sns.kdeplot(train.loc[~disaster]['unique word count'], shade = True, label = 'Not Disaster', ax=axes[0]) graph1 = sns.kdeplot(train.loc[disaster]['unique word count'], shade = True, label = 'Disaster', ax=axes[0]) graph1.set_title('Distribution of Unique Word Count') graph2 = sns.kdeplot(train.loc[~disaster]['stopword count'], shade = True, label = 'Not Disaster', ax=axes[1]) graph2 = sns.kdeplot(train.loc[disaster]['stopword count'], shade = True, label = 'Disaster', ax=axes[1]) graph2.set_title('Distribution of Stopword Word Count') graph3 = sns.kdeplot(train.loc[~disaster]['punctuation count'], shade = True, label = 'Not Disaster', ax=axes[2], bw = 1) graph3 = sns.kdeplot(train.loc[disaster]['punctuation count'], shade = True, label = 'Disaster', ax=axes[2], bw = 1) graph3.set_title('Distribution of Punctuation Count') graph4 = sns.kdeplot(train.loc[~disaster]['stopword ratio'], shade = True, label = 'Not Disaster', ax=axes[3], bw =.05) graph4 = sns.kdeplot(train.loc[disaster]['stopword ratio'], shade = True, label = 'Disaster', ax=axes[3], bw =.05) graph4.set_title('Distribution of Stopword Ratio') fig.tight_layout() plt.show()<string_transform>
class_survived = [col for col in probs.columns if col.endswith('Prob_1')] probs['1'] = probs[class_survived].sum(axis=1)/ N probs['0'] = probs.drop(columns=class_survived ).sum(axis=1)/ N probs['pred'] = 0 pos = probs[probs['1'] >= 0.5].index probs.loc[pos, 'pred'] = 1 y_pred = probs['pred'].astype(int) submission_df = pd.DataFrame(columns=['PassengerId', 'Survived']) submission_df['PassengerId'] = df_test['PassengerId'] submission_df['Survived'] = y_pred.values submission_df.to_csv('submissions.csv', header=True, index=False) submission_df.head(10 )
Titanic - Machine Learning from Disaster
11,548,643
def remove_punctuation(x): return x.translate(str.maketrans('', '', string.punctuation)) def remove_stopwords(x): return ' '.join([i for i in x.split() if i not in wordcloud.STOPWORDS]) def remove_less_than(x): return ' '.join([i for i in x.split() if len(i)> 3]) def remove_non_alphabet(x): return ' '.join([i for i in x.split() if i.isalpha() ]) def strip_all_entities(x): return ' '.join(re.sub("(@[A-Za-z0-9]+)|([^0-9A-Za-z \t])|(\w+:\/\/\S+)"," ",x ).split() )<drop_column>
!pip install fastai2
Titanic - Machine Learning from Disaster
11,548,643
strip_all_entities('@shawn Titanic Times: Telegraph.co.ukTitanic tragedy could have been preve...http://bet.ly/tuN2wx' )<string_transform>
!pip install fastcore==0.1.35
Titanic - Machine Learning from Disaster
11,548,643
!pip install autocorrect def spell_check(x): spell = Speller(lang='en') return " ".join([spell(i)for i in x.split() ]) mispelled = 'Pleaze spelcheck this sentince' spell_check(mispelled )<feature_engineering>
fastcore.__version__
Titanic - Machine Learning from Disaster
11,548,643
PROCESS_TWEETS = False if PROCESS_TWEETS: total['text'] = total['text'].apply(lambda x: x.lower()) total['text'] = total['text'].apply(lambda x: re.sub(r'https?://\S+|www\.\S+', '', x, flags = re.MULTILINE)) total['text'] = total['text'].apply(remove_punctuation) total['text'] = total['text'].apply(remove_stopwords) total['text'] = total['text'].apply(remove_less_than) total['text'] = total['text'].apply(remove_non_alphabet) total['text'] = total['text'].apply(spell_check )<define_variables>
fastai2.__version__
Titanic - Machine Learning from Disaster
11,548,643
contractions = { "ain't": "am not / are not / is not / has not / have not", "aren't": "are not / am not", "can't": "cannot", "can't've": "cannot have", "'cause": "because", "could've": "could have", "couldn't": "could not", "couldn't've": "could not have", "didn't": "did not", "doesn't": "does not", "don't": "do not", "hadn't": "had not", "hadn't've": "had not have", "hasn't": "has not", "haven't": "have not", "he'd": "he had / he would", "he'd've": "he would have", "he'll": "he shall / he will", "he'll've": "he shall have / he will have", "he's": "he has / he is", "how'd": "how did", "how'd'y": "how do you", "how'll": "how will", "how's": "how has / how is / how does", "I'd": "I had / I would", "I'd've": "I would have", "I'll": "I shall / I will", "I'll've": "I shall have / I will have", "I'm": "I am", "I've": "I have", "isn't": "is not", "it'd": "it had / it would", "it'd've": "it would have", "it'll": "it shall / it will", "it'll've": "it shall have / it will have", "it's": "it has / it is", "let's": "let us", "ma'am": "madam", "mayn't": "may not", "might've": "might have", "mightn't": "might not", "mightn't've": "might not have", "must've": "must have", "mustn't": "must not", "mustn't've": "must not have", "needn't": "need not", "needn't've": "need not have", "o'clock": "of the clock", "oughtn't": "ought not", "oughtn't've": "ought not have", "shan't": "shall not", "sha'n't": "shall not", "shan't've": "shall not have", "she'd": "she had / she would", "she'd've": "she would have", "she'll": "she shall / she will", "she'll've": "she shall have / she will have", "she's": "she has / she is", "should've": "should have", "shouldn't": "should not", "shouldn't've": "should not have", "so've": "so have", "so's": "so as / so is", "that'd": "that would / that had", "that'd've": "that would have", "that's": "that has / that is", "there'd": "there had / there would", "there'd've": "there would have", "there's": "there has / there is", "they'd": "they had / they would", "they'd've": "they would have", "they'll": "they shall / they will", "they'll've": "they shall have / they will have", "they're": "they are", "they've": "they have", "to've": "to have", "wasn't": "was not", "we'd": "we had / we would", "we'd've": "we would have", "we'll": "we will", "we'll've": "we will have", "we're": "we are", "we've": "we have", "weren't": "were not", "what'll": "what shall / what will", "what'll've": "what shall have / what will have", "what're": "what are", "what's": "what has / what is", "what've": "what have", "when's": "when has / when is", "when've": "when have", "where'd": "where did", "where's": "where has / where is", "where've": "where have", "who'll": "who shall / who will", "who'll've": "who shall have / who will have", "who's": "who has / who is", "who've": "who have", "why's": "why has / why is", "why've": "why have", "will've": "will have", "won't": "will not", "won't've": "will not have", "would've": "would have", "wouldn't": "would not", "wouldn't've": "would not have", "y'all": "you all", "y'all'd": "you all would", "y'all'd've": "you all would have", "y'all're": "you all are", "y'all've": "you all have", "you'd": "you had / you would", "you'd've": "you would have", "you'll": "you shall / you will", "you'll've": "you shall have / you will have", "you're": "you are", "you've": "you have" } contractions_re = re.compile('(%s)' % '|'.join(contractions.keys())) def expand_contractions(s, contractions = contractions): def replace(match): return contractions[match.group(0)] return contractions_re.sub(replace, s) expand_contractions("can't stop won't stop" )<feature_engineering>
from fastai2.tabular.all import *
Titanic - Machine Learning from Disaster
11,548,643
total['text'] = total['text'].apply(expand_contractions )<categorify>
df_test= pd.read_csv('/kaggle/input/titanic-extended/test.csv') df_train= pd.read_csv('.. /input/titanic-extended/train.csv') df_train.head()
Titanic - Machine Learning from Disaster
11,548,643
def clean(tweet): tweet = re.sub(r"tnwx", "Tennessee Weather", tweet) tweet = re.sub(r"azwx", "Arizona Weather", tweet) tweet = re.sub(r"alwx", "Alabama Weather", tweet) tweet = re.sub(r"wordpressdotcom", "wordpress", tweet) tweet = re.sub(r"gawx", "Georgia Weather", tweet) tweet = re.sub(r"scwx", "South Carolina Weather", tweet) tweet = re.sub(r"cawx", "California Weather", tweet) tweet = re.sub(r"usNWSgov", "United States National Weather Service", tweet) tweet = re.sub(r"MH370", "Malaysia Airlines Flight 370", tweet) tweet = re.sub(r"okwx", "Oklahoma City Weather", tweet) tweet = re.sub(r"arwx", "Arkansas Weather", tweet) tweet = re.sub(r"lmao", "laughing my ass off", tweet) tweet = re.sub(r"amirite", "am I right", tweet) tweet = re.sub(r"w/e", "whatever", tweet) tweet = re.sub(r"w/", "with", tweet) tweet = re.sub(r"USAgov", "USA government", tweet) tweet = re.sub(r"recentlu", "recently", tweet) tweet = re.sub(r"Ph0tos", "Photos", tweet) tweet = re.sub(r"exp0sed", "exposed", tweet) tweet = re.sub(r"<3", "love", tweet) tweet = re.sub(r"amageddon", "armageddon", tweet) tweet = re.sub(r"Trfc", "Traffic", tweet) tweet = re.sub(r"WindStorm", "Wind Storm", tweet) tweet = re.sub(r"16yr", "16 year", tweet) tweet = re.sub(r"TRAUMATISED", "traumatized", tweet) tweet = re.sub(r"IranDeal", "Iran Deal", tweet) tweet = re.sub(r"ArianaGrande", "Ariana Grande", tweet) tweet = re.sub(r"camilacabello97", "camila cabello", tweet) tweet = re.sub(r"RondaRousey", "Ronda Rousey", tweet) tweet = re.sub(r"MTVHottest", "MTV Hottest", tweet) tweet = re.sub(r"TrapMusic", "Trap Music", tweet) tweet = re.sub(r"ProphetMuhammad", "Prophet Muhammad", tweet) tweet = re.sub(r"PantherAttack", "Panther Attack", tweet) tweet = re.sub(r"StrategicPatience", "Strategic Patience", tweet) tweet = re.sub(r"socialnews", "social news", tweet) tweet = re.sub(r"IDPs:", "Internally Displaced People :", tweet) tweet = re.sub(r"ArtistsUnited", "Artists United", tweet) tweet = re.sub(r"ClaytonBryant", "Clayton Bryant", tweet) tweet = re.sub(r"jimmyfallon", "jimmy fallon", tweet) tweet = re.sub(r"justinbieber", "justin bieber", tweet) tweet = re.sub(r"Time2015", "Time 2015", tweet) tweet = re.sub(r"djicemoon", "dj icemoon", tweet) tweet = re.sub(r"LivingSafely", "Living Safely", tweet) tweet = re.sub(r"FIFA16", "Fifa 2016", tweet) tweet = re.sub(r"thisiswhywecanthavenicethings", "this is why we cannot have nice things", tweet) tweet = re.sub(r"bbcnews", "bbc news", tweet) tweet = re.sub(r"UndergroundRailraod", "Underground Railraod", tweet) tweet = re.sub(r"c4news", "c4 news", tweet) tweet = re.sub(r"MUDSLIDE", "mudslide", tweet) tweet = re.sub(r"NoSurrender", "No Surrender", tweet) tweet = re.sub(r"NotExplained", "Not Explained", tweet) tweet = re.sub(r"greatbritishbakeoff", "great british bake off", tweet) tweet = re.sub(r"LondonFire", "London Fire", tweet) tweet = re.sub(r"KOTAWeather", "KOTA Weather", tweet) tweet = re.sub(r"LuchaUnderground", "Lucha Underground", tweet) tweet = re.sub(r"KOIN6News", "KOIN 6 News", tweet) tweet = re.sub(r"LiveOnK2", "Live On K2", tweet) tweet = re.sub(r"9NewsGoldCoast", "9 News Gold Coast", tweet) tweet = re.sub(r"nikeplus", "nike plus", tweet) tweet = re.sub(r"david_cameron", "David Cameron", tweet) tweet = re.sub(r"peterjukes", "Peter Jukes", tweet) tweet = re.sub(r"MikeParrActor", "Michael Parr", tweet) tweet = re.sub(r"4PlayThursdays", "Foreplay Thursdays", tweet) tweet = re.sub(r"TGF2015", "Tontitown Grape Festival", tweet) tweet = re.sub(r"realmandyrain", "Mandy Rain", tweet) tweet = re.sub(r"GraysonDolan", "Grayson Dolan", tweet) tweet = re.sub(r"ApolloBrown", "Apollo Brown", tweet) tweet = re.sub(r"saddlebrooke", "Saddlebrooke", tweet) tweet = re.sub(r"TontitownGrape", "Tontitown Grape", tweet) tweet = re.sub(r"AbbsWinston", "Abbs Winston", tweet) tweet = re.sub(r"ShaunKing", "Shaun King", tweet) tweet = re.sub(r"MeekMill", "Meek Mill", tweet) tweet = re.sub(r"TornadoGiveaway", "Tornado Giveaway", tweet) tweet = re.sub(r"GRupdates", "GR updates", tweet) tweet = re.sub(r"SouthDowns", "South Downs", tweet) tweet = re.sub(r"braininjury", "brain injury", tweet) tweet = re.sub(r"auspol", "Australian politics", tweet) tweet = re.sub(r"PlannedParenthood", "Planned Parenthood", tweet) tweet = re.sub(r"calgaryweather", "Calgary Weather", tweet) tweet = re.sub(r"weallheartonedirection", "we all heart one direction", tweet) tweet = re.sub(r"edsheeran", "Ed Sheeran", tweet) tweet = re.sub(r"TrueHeroes", "True Heroes", tweet) tweet = re.sub(r"ComplexMag", "Complex Magazine", tweet) tweet = re.sub(r"TheAdvocateMag", "The Advocate Magazine", tweet) tweet = re.sub(r"CityofCalgary", "City of Calgary", tweet) tweet = re.sub(r"EbolaOutbreak", "Ebola Outbreak", tweet) tweet = re.sub(r"SummerFate", "Summer Fate", tweet) tweet = re.sub(r"RAmag", "Royal Academy Magazine", tweet) tweet = re.sub(r"offers2go", "offers to go", tweet) tweet = re.sub(r"ModiMinistry", "Modi Ministry", tweet) tweet = re.sub(r"TAXIWAYS", "taxi ways", tweet) tweet = re.sub(r"Calum5SOS", "Calum Hood", tweet) tweet = re.sub(r"JamesMelville", "James Melville", tweet) tweet = re.sub(r"JamaicaObserver", "Jamaica Observer", tweet) tweet = re.sub(r"TweetLikeItsSeptember11th2001", "Tweet like it is september 11th 2001", tweet) tweet = re.sub(r"cbplawyers", "cbp lawyers", tweet) tweet = re.sub(r"fewmoretweets", "few more tweets", tweet) tweet = re.sub(r"BlackLivesMatter", "Black Lives Matter", tweet) tweet = re.sub(r"NASAHurricane", "NASA Hurricane", tweet) tweet = re.sub(r"onlinecommunities", "online communities", tweet) tweet = re.sub(r"humanconsumption", "human consumption", tweet) tweet = re.sub(r"Typhoon-Devastated", "Typhoon Devastated", tweet) tweet = re.sub(r"Meat-Loving", "Meat Loving", tweet) tweet = re.sub(r"facialabuse", "facial abuse", tweet) tweet = re.sub(r"LakeCounty", "Lake County", tweet) tweet = re.sub(r"BeingAuthor", "Being Author", tweet) tweet = re.sub(r"withheavenly", "with heavenly", tweet) tweet = re.sub(r"thankU", "thank you", tweet) tweet = re.sub(r"iTunesMusic", "iTunes Music", tweet) tweet = re.sub(r"OffensiveContent", "Offensive Content", tweet) tweet = re.sub(r"WorstSummerJob", "Worst Summer Job", tweet) tweet = re.sub(r"HarryBeCareful", "Harry Be Careful", tweet) tweet = re.sub(r"NASASolarSystem", "NASA Solar System", tweet) tweet = re.sub(r"animalrescue", "animal rescue", tweet) tweet = re.sub(r"KurtSchlichter", "Kurt Schlichter", tweet) tweet = re.sub(r"Throwingknifes", "Throwing knives", tweet) tweet = re.sub(r"GodsLove", "God's Love", tweet) tweet = re.sub(r"bookboost", "book boost", tweet) tweet = re.sub(r"ibooklove", "I book love", tweet) tweet = re.sub(r"NestleIndia", "Nestle India", tweet) tweet = re.sub(r"realDonaldTrump", "Donald Trump", tweet) tweet = re.sub(r"DavidVonderhaar", "David Vonderhaar", tweet) tweet = re.sub(r"CecilTheLion", "Cecil The Lion", tweet) tweet = re.sub(r"weathernetwork", "weather network", tweet) tweet = re.sub(r"GOPDebate", "GOP Debate", tweet) tweet = re.sub(r"RickPerry", "Rick Perry", tweet) tweet = re.sub(r"frontpage", "front page", tweet) tweet = re.sub(r"NewsInTweets", "News In Tweets", tweet) tweet = re.sub(r"ViralSpell", "Viral Spell", tweet) tweet = re.sub(r"til_now", "until now", tweet) tweet = re.sub(r"volcanoinRussia", "volcano in Russia", tweet) tweet = re.sub(r"ZippedNews", "Zipped News", tweet) tweet = re.sub(r"MicheleBachman", "Michele Bachman", tweet) tweet = re.sub(r"53inch", "53 inch", tweet) tweet = re.sub(r"KerrickTrial", "Kerrick Trial", tweet) tweet = re.sub(r"abstorm", "Alberta Storm", tweet) tweet = re.sub(r"Beyhive", "Beyonce hive", tweet) tweet = re.sub(r"RockyFire", "Rocky Fire", tweet) tweet = re.sub(r"Listen/Buy", "Listen / Buy", tweet) tweet = re.sub(r"ArtistsUnited", "Artists United", tweet) tweet = re.sub(r"ENGvAUS", "England vs Australia", tweet) tweet = re.sub(r"ScottWalker", "Scott Walker", tweet) return tweet total['text'] = total['text'].apply(clean )<define_variables>
df_train.isnull().sum().sort_index() /len(df_train )
Titanic - Machine Learning from Disaster
11,548,643
tweets = [tweet for tweet in total['text']] train = total[:len(train)] test = total[len(train):]<categorify>
df_train.dtypes g_train =df_train.columns.to_series().groupby(df_train.dtypes ).groups g_train
Titanic - Machine Learning from Disaster
11,548,643
def generate_ngrams(text, n_gram=1): token = [token for token in text.lower().split(' ')if token != '' if token not in wordcloud.STOPWORDS] ngrams = zip(*[token[i:] for i in range(n_gram)]) return [' '.join(ngram)for ngram in ngrams] disaster_unigrams = defaultdict(int) for word in total[train['target'] == 1]['text']: for word in generate_ngrams(word, n_gram = 1): disaster_unigrams[word] += 1 disaster_unigrams = pd.DataFrame(sorted(disaster_unigrams.items() , key=lambda x: x[1])[::-1]) nondisaster_unigrams = defaultdict(int) for word in total[train['target'] == 0]['text']: for word in generate_ngrams(word, n_gram = 1): nondisaster_unigrams[word] += 1 nondisaster_unigrams = pd.DataFrame(sorted(nondisaster_unigrams.items() , key=lambda x: x[1])[::-1]) disaster_bigrams = defaultdict(int) for word in total[train['target'] == 1]['text']: for word in generate_ngrams(word, n_gram = 2): disaster_bigrams[word] += 1 disaster_bigrams = pd.DataFrame(sorted(disaster_bigrams.items() , key=lambda x: x[1])[::-1]) nondisaster_bigrams = defaultdict(int) for word in total[train['target'] == 0]['text']: for word in generate_ngrams(word, n_gram = 2): nondisaster_bigrams[word] += 1 nondisaster_bigrams = pd.DataFrame(sorted(nondisaster_bigrams.items() , key=lambda x: x[1])[::-1]) disaster_trigrams = defaultdict(int) for word in total[train['target'] == 1]['text']: for word in generate_ngrams(word, n_gram = 3): disaster_trigrams[word] += 1 disaster_trigrams = pd.DataFrame(sorted(disaster_trigrams.items() , key=lambda x: x[1])[::-1]) nondisaster_trigrams = defaultdict(int) for word in total[train['target'] == 0]['text']: for word in generate_ngrams(word, n_gram = 3): nondisaster_trigrams[word] += 1 nondisaster_trigrams = pd.DataFrame(sorted(nondisaster_trigrams.items() , key=lambda x: x[1])[::-1]) disaster_4grams = defaultdict(int) for word in total[train['target'] == 1]['text']: for word in generate_ngrams(word, n_gram = 4): disaster_4grams[word] += 1 disaster_4grams = pd.DataFrame(sorted(disaster_4grams.items() , key=lambda x: x[1])[::-1]) nondisaster_4grams = defaultdict(int) for word in total[train['target'] == 0]['text']: for word in generate_ngrams(word, n_gram = 4): nondisaster_4grams[word] += 1 nondisaster_4grams = pd.DataFrame(sorted(nondisaster_4grams.items() , key=lambda x: x[1])[::-1] )<string_transform>
cat_names= [ 'Name', 'Sex', 'Ticket', 'Cabin', 'Embarked', 'Name_wiki', 'Hometown', 'Boarded', 'Destination', 'Lifeboat', 'Body' ] cont_names = [ 'PassengerId', 'Pclass', 'SibSp', 'Parch', 'Age', 'Fare', 'WikiId', 'Age_wiki','Class' ]
Titanic - Machine Learning from Disaster
11,548,643
to_exclude = '*+-/() % [\\]{|}^_`~\t' to_tokenize = '!" tokenizer = Tokenizer(filters = to_exclude) text = 'Why are you so f% text = re.sub(r'(['+to_tokenize+'])', r' \1 ', text) tokenizer.fit_on_texts([text]) print(tokenizer.word_index )<feature_engineering>
splits = RandomSplitter(valid_pct=0.2 )(range_of(df_train)) to = TabularPandas(df_train, procs=[Categorify, FillMissing,Normalize], cat_names = cat_names, cont_names = cont_names, y_names='Survived', splits=splits )
Titanic - Machine Learning from Disaster
11,548,643
<string_transform>
g_train =to.train.xs.columns.to_series().groupby(to.train.xs.dtypes ).groups g_train
Titanic - Machine Learning from Disaster
11,548,643
tokenizer = Tokenizer() tokenizer.fit_on_texts(tweets) sequences = tokenizer.texts_to_sequences(tweets) word_index = tokenizer.word_index print('Found %s unique tokens.' % len(word_index)) data = pad_sequences(sequences) labels = train['target'] print('Shape of data tensor:', data.shape) print('Shape of label tensor:', labels.shape) nlp_train = data[:len(train)] labels = labels nlp_test = data[len(train):] MAX_SEQUENCE_LENGTH = data.shape[1]<feature_engineering>
to.train
Titanic - Machine Learning from Disaster
11,548,643
embeddings_index = {} with open('.. /input/glove-global-vectors-for-word-representation/glove.6B.200d.txt','r')as f: for line in tqdm(f): values = line.split() word = values[0] coefs = np.asarray(values[1:], dtype='float32') embeddings_index[word] = coefs f.close() print('Found %s word vectors in the GloVe library' % len(embeddings_index))<define_variables>
to.train.xs
Titanic - Machine Learning from Disaster
11,548,643
EMBEDDING_DIM = 200<categorify>
X_train, y_train = to.train.xs, to.train.ys.values.ravel() X_valid, y_valid = to.valid.xs, to.valid.ys.values.ravel()
Titanic - Machine Learning from Disaster
11,548,643
embedding_matrix = np.zeros(( len(word_index)+ 1, EMBEDDING_DIM)) for word, i in tqdm(word_index.items()): embedding_vector = embeddings_index.get(word) if embedding_vector is not None: embedding_matrix[i] = embedding_vector print("Our embedded matrix is of dimension", embedding_matrix.shape )<choose_model_class>
rnf_classifier= RandomForestClassifier(n_estimators=100, n_jobs=-1) rnf_classifier.fit(X_train,y_train )
Titanic - Machine Learning from Disaster
11,548,643
embedding = Embedding(len(word_index)+ 1, EMBEDDING_DIM, weights = [embedding_matrix], input_length = MAX_SEQUENCE_LENGTH, trainable = False) <normalization>
y_pred=rnf_classifier.predict(X_valid) accuracy_score(y_pred, y_valid )
Titanic - Machine Learning from Disaster
11,548,643
def scale(df, scaler): return scaler.fit_transform(df.iloc[:, 2:]) meta_train = scale(train, StandardScaler()) meta_test = scale(test, StandardScaler() )<choose_model_class>
df_test.dtypes g_train =df_test.columns.to_series().groupby(df_test.dtypes ).groups g_train
Titanic - Machine Learning from Disaster
11,548,643
def create_lstm(spatial_dropout, dropout, recurrent_dropout, learning_rate, bidirectional = False): activation = LeakyReLU(alpha = 0.01) nlp_input = Input(shape =(MAX_SEQUENCE_LENGTH,), name = 'nlp_input') meta_input_train = Input(shape =(7,), name = 'meta_train') emb = embedding(nlp_input) emb = SpatialDropout1D(dropout )(emb) if bidirectional: nlp_out =(Bidirectional(LSTM(100, dropout = dropout, recurrent_dropout = recurrent_dropout, kernel_initializer = 'orthogonal')) )(emb) else: nlp_out =(LSTM(100, dropout = dropout, recurrent_dropout = recurrent_dropout, kernel_initializer = 'orthogonal'))(emb) x = Concatenate()([nlp_out, meta_input_train]) x = Dropout(dropout )(x) preds = Dense(1, activation='sigmoid', kernel_regularizer = regularizers.l2(1e-4))(x) model = Model(inputs=[nlp_input , meta_input_train], outputs = preds) optimizer = Adam(learning_rate = learning_rate) model.compile(loss = 'binary_crossentropy', optimizer = optimizer, metrics = ['accuracy']) return model<choose_model_class>
cat_names= [ 'Name', 'Sex', 'Ticket', 'Cabin', 'Embarked', 'Name_wiki', 'Hometown', 'Boarded', 'Destination', 'Lifeboat', 'Body' ] cont_names = [ 'PassengerId', 'Pclass', 'SibSp', 'Parch', 'Age', 'Fare', 'WikiId', 'Age_wiki','Class' ]
Titanic - Machine Learning from Disaster
11,548,643
lstm = create_lstm(spatial_dropout =.2, dropout =.2, recurrent_dropout =.2, learning_rate = 3e-4, bidirectional = True) lstm.summary()<train_model>
test = TabularPandas(df_test, procs=[Categorify, FillMissing,Normalize], cat_names = cat_names, cont_names = cont_names, )
Titanic - Machine Learning from Disaster
11,548,643
history1 = lstm.fit([nlp_train, meta_train], labels, validation_split =.2, epochs = 5, batch_size = 21, verbose = 1 )<choose_model_class>
X_test= test.train.xs
Titanic - Machine Learning from Disaster
11,548,643
callback = EarlyStopping(monitor = 'val_loss', patience = 4) <choose_model_class>
X_test.dtypes g_train =X_test.columns.to_series().groupby(X_test.dtypes ).groups g_train
Titanic - Machine Learning from Disaster
11,548,643
def create_lstm_2(spatial_dropout, dropout, recurrent_dropout, learning_rate, bidirectional = False): activation = LeakyReLU(alpha = 0.01) nlp_input = Input(shape =(MAX_SEQUENCE_LENGTH,), name = 'nlp_input') meta_input_train = Input(shape =(7,), name = 'meta_train') emb = embedding(nlp_input) emb = SpatialDropout1D(dropout )(emb) if bidirectional: nlp_out =(Bidirectional(LSTM(100, dropout = dropout, recurrent_dropout = recurrent_dropout, kernel_initializer = 'orthogonal')) )(emb) else: nlp_out =(LSTM(100, dropout = dropout, recurrent_dropout = recurrent_dropout, kernel_initializer = 'orthogonal'))(emb) x = Concatenate()([nlp_out, meta_input_train]) x = Dropout(dropout )(x) x =(Dense(100, activation = activation, kernel_regularizer = regularizers.l2(1e-4), kernel_initializer = 'he_normal'))(x) x = Dropout(dropout )(x) preds = Dense(1, activation='sigmoid', kernel_regularizer = regularizers.l2(1e-4))(x) model = Model(inputs=[nlp_input , meta_input_train], outputs = preds) optimizer = Adam(learning_rate = learning_rate) model.compile(loss = 'binary_crossentropy', optimizer = optimizer, metrics = ['accuracy']) return model<choose_model_class>
X_test= X_test.drop('Fare_na', axis=1 )
Titanic - Machine Learning from Disaster
11,548,643
lstm_2 = create_lstm_2(spatial_dropout =.4, dropout =.4, recurrent_dropout =.4, learning_rate = 3e-4, bidirectional = True) lstm_2.summary()<train_model>
y_pred=rnf_classifier.predict(X_test)
Titanic - Machine Learning from Disaster
11,548,643
history2 = lstm_2.fit([nlp_train, meta_train], labels, validation_split =.2, epochs = 30, batch_size = 21, verbose = 1 )<predict_on_test>
y_pred= y_pred.astype(int )
Titanic - Machine Learning from Disaster
11,548,643
<choose_model_class><EOS>
output= pd.DataFrame({'PassengerId':df_test.PassengerId, 'Survived': y_pred}) output.to_csv('my_submission_titanic.csv', index=False) output.head()
Titanic - Machine Learning from Disaster
4,005,064
<SOS> metric: categorizationaccuracy Kaggle data source: titanic-machine-learning-from-disaster<train_model>
print(os.listdir(".. /input"))
Titanic - Machine Learning from Disaster
4,005,064
history3 = dual_lstm.fit([nlp_train, meta_train], labels, validation_split =.2, epochs = 25, batch_size = 21, verbose = 1 )<predict_on_test>
train_df = pd.read_csv(".. /input/train.csv") test_df = pd.read_csv(".. /input/test.csv") print("Train shape : ",train_df.shape) print("Test shape : ",test_df.shape )
Titanic - Machine Learning from Disaster
4,005,064
submission_lstm2 = pd.DataFrame() submission_lstm2['id'] = test_id submission_lstm2['prob'] = dual_lstm.predict([nlp_test, meta_test]) submission_lstm2['target'] = submission_lstm2['prob'].apply(lambda x: 0 if x <.5 else 1) submission_lstm2.head(10 )<define_variables>
tabla_completa = train_df.append(test_df, ignore_index=True )
Titanic - Machine Learning from Disaster
4,005,064
BATCH_SIZE = 32 EPOCHS = 2 USE_META = True ADD_DENSE = False DENSE_DIM = 64 ADD_DROPOUT = False DROPOUT =.2<install_modules>
tabla_completa["Cabin"] = tabla_completa["Cabin"].fillna("N") letras_cabinas = tabla_completa["Cabin"].str[0].unique().tolist()
Titanic - Machine Learning from Disaster
4,005,064
!pip install --quiet transformers <categorify>
tabla_completa["Letra_Cabina"] = tabla_completa["Cabin"].str[0]
Titanic - Machine Learning from Disaster
4,005,064
TOKENIZER = AutoTokenizer.from_pretrained("bert-large-uncased") enc = TOKENIZER.encode("Encode me!") dec = TOKENIZER.decode(enc) print("Encode: " + str(enc)) print("Decode: " + str(dec))<categorify>
dic_titulos_simples = { "Mr" : "Mr", "Mrs" : "Mrs", "Miss" : "Miss", "Master" : "Master", "Don" : "Nobleza", "Rev" : "Oficial", "Dr" : "Oficial", "Mme" : "Mrs", "Ms" : "Mrs", "Major" : "Oficial", "Lady" : "Nobleza", "Sir" : "Nobleza", "Mlle" : "Miss", "Col" : "Oficial", "Capt" : "Oficial", "Countess" : "Nobleza", "Jonkheer" : "Nobleza", "Dona" : "Nobleza" } tabla_completa["Titulo"] = tabla_completa["Name"].str.partition('.')[0].str.rpartition() [2].map(dic_titulos_simples )
Titanic - Machine Learning from Disaster
4,005,064
def bert_encode(data,maximum_len): input_ids = [] attention_masks = [] for i in range(len(data.text)) : encoded = TOKENIZER.encode_plus(data.text[i], add_special_tokens=True, max_length=maximum_len, pad_to_max_length=True, return_attention_mask=True) input_ids.append(encoded['input_ids']) attention_masks.append(encoded['attention_mask']) return np.array(input_ids),np.array(attention_masks )<choose_model_class>
grupo = tabla_completa.groupby(['Sex','Pclass','Titulo']) tabla_completa['Age'] = grupo['Age'].apply(lambda x: x.fillna(x.median()))
Titanic - Machine Learning from Disaster
4,005,064
def build_model(model_layer, learning_rate, use_meta = USE_META, add_dense = ADD_DENSE, dense_dim = DENSE_DIM, add_dropout = ADD_DROPOUT, dropout = DROPOUT): input_ids = tf.keras.Input(shape=(60,),dtype='int32') attention_masks = tf.keras.Input(shape=(60,),dtype='int32') meta_input = tf.keras.Input(shape =(meta_train.shape[1],)) transformer_layer = model_layer([input_ids,attention_masks]) output = transformer_layer[1] if use_meta: output = tf.keras.layers.Concatenate()([output, meta_input]) if add_dense: print("Training with additional dense layer...") output = tf.keras.layers.Dense(dense_dim,activation='relu' )(output) if add_dropout: print("Training with dropout...") output = tf.keras.layers.Dropout(dropout )(output) output = tf.keras.layers.Dense(1,activation='sigmoid' )(output) if use_meta: print("Training with meta-data...") model = tf.keras.models.Model(inputs = [input_ids,attention_masks, meta_input],outputs = output) else: print("Training without meta-data...") model = tf.keras.models.Model(inputs = [input_ids,attention_masks],outputs = output) model.compile(tf.keras.optimizers.Adam(lr=learning_rate), loss='binary_crossentropy', metrics=['accuracy']) return model<load_from_csv>
tabla_completa['Fare'] = tabla_completa['Fare'].fillna(tabla_completa['Fare'].median() )
Titanic - Machine Learning from Disaster
4,005,064
train = pd.read_csv('.. /input/nlp-getting-started/train.csv') test = pd.read_csv('.. /input/nlp-getting-started/test.csv' )<categorify>
tabla_completa['Num_Familiares'] = tabla_completa['SibSp'] + tabla_completa['Parch']
Titanic - Machine Learning from Disaster
4,005,064
bert_large = TFAutoModel.from_pretrained('bert-large-uncased') TOKENIZER = AutoTokenizer.from_pretrained("bert-large-uncased") train_input_ids,train_attention_masks = bert_encode(train,60) test_input_ids,test_attention_masks = bert_encode(test,60) print('Train length:', len(train_input_ids)) print('Test length:', len(test_input_ids)) BERT_large = build_model(bert_large, learning_rate = 1e-5) BERT_large.summary()<train_model>
tabla_completa['Sex'] = tabla_completa['Sex'].map({"male": 0, "female":1}) dummies_pclass = pd.get_dummies(tabla_completa['Pclass'], prefix="Pclass") dummies_titulo = pd.get_dummies(tabla_completa['Titulo'], prefix="Titulo") dummies_letra_cab = pd.get_dummies(tabla_completa['Letra_Cabina'], prefix="Letra_Cabina") dummies_grupo_edad = pd.get_dummies(tabla_completa['Grupo_Edad'], prefix="Grupo_Edad") dummies_tabla = pd.concat([tabla_completa, dummies_pclass, dummies_titulo, dummies_letra_cab,dummies_grupo_edad], axis=1) dummies_tabla.drop(['Pclass', 'Titulo', 'Cabin', 'Letra_Cabina', 'Grupo_Edad', 'Embarked', 'Name', 'Ticket'], axis=1, inplace=True )
Titanic - Machine Learning from Disaster
4,005,064
history_bert = BERT_large.fit([train_input_ids,train_attention_masks, meta_train], train.target, validation_split =.2, epochs = EPOCHS, callbacks = [checkpoint], batch_size = BATCH_SIZE )<predict_on_test>
train_df = dummies_tabla[ :len(train_df)] test_df = dummies_tabla[(len(dummies_tabla)- len(test_df)) : ] train_df['Survived'] = train_df['Survived'].astype(int) print("Train: ",train_df) print("Test: ",test_df )
Titanic - Machine Learning from Disaster
4,005,064
BERT_large.load_weights('large_model.h5') preds_bert = BERT_large.predict([test_input_ids,test_attention_masks,meta_test] )<prepare_output>
test_df = test_df.reset_index(drop = True) train_df = train_df.reset_index(drop = True) train_dfX = train_df.drop(['PassengerId','Survived'], axis=1) train_dfY = train_df['Survived'] submission = pd.DataFrame(data=test_df['PassengerId'].copy()) print(submission) test_df = test_df.drop(['PassengerId', 'Survived'], axis=1 )
Titanic - Machine Learning from Disaster
4,005,064
submission_bert = pd.DataFrame() submission_bert['id'] = test_id submission_bert['prob'] = preds_bert submission_bert['target'] = np.round(submission_bert['prob'] ).astype(int) submission_bert.head(10 )<save_to_csv>
sc = StandardScaler() train_dfX = sc.fit_transform(train_dfX) test_df = sc.transform(test_df )
Titanic - Machine Learning from Disaster
4,005,064
submission_bert = submission_bert[['id', 'target']] submission_bert.to_csv('submission_bert.csv', index = False) print('Blended submission has been saved to disk' )<set_options>
train_dfX,val_dfX,train_dfY, val_dfY = train_test_split(train_dfX,train_dfY , test_size=0.1, stratify=train_dfY) print("Entrnamiento: ",train_dfX.shape) print("Validacion : ",val_dfX.shape )
Titanic - Machine Learning from Disaster
4,005,064
plt.style.use('ggplot') warnings.filterwarnings('ignore') <define_variables>
def func_model(arquitectura): np.random.seed(42) random_seed = 42 first =True inp = Input(shape=(train_dfX.shape[1],)) for capa in arquitectura: if first: x=Dense(capa, activation="relu", kernel_initializer=initializers.RandomNormal(seed=random_seed), bias_initializer='zeros' )(inp) first = False else: x=Dense(capa, activation="relu", kernel_initializer=initializers.RandomNormal(seed=random_seed), bias_initializer='zeros' )(x) x=Dense(1, activation="sigmoid", kernel_initializer=initializers.RandomNormal(seed=random_seed), bias_initializer='zeros' )(x) model = Model(inputs=inp, outputs=x) model.compile(loss='binary_crossentropy', optimizer=optimizers.Adam(lr=0.0002), metrics=['binary_accuracy']) return model
Titanic - Machine Learning from Disaster
4,005,064
def seed_everything(seed): os.environ['PYTHONHASHSEED']=str(seed) tf.random.set_seed(seed) np.random.seed(seed) random.seed(seed) seed_everything(34 )<load_from_csv>
arq1 = [1024, 1024, 512] model1 = None model1 = func_model(arq1) train_history_tam1 = model1.fit(train_dfX, train_dfY, batch_size=32, epochs=epochs, validation_data=(val_dfX, val_dfY)) graf_model(train_history_tam1) precision(model1 )
Titanic - Machine Learning from Disaster
4,005,064
train = pd.read_csv('.. /input/nlp-getting-started/train.csv') test = pd.read_csv('.. /input/nlp-getting-started/test.csv') train.head()<categorify>
arq2 = [1024, 512, 512] model2 = None model2 = func_model(arq2) train_history_tam2 = model2.fit(train_dfX, train_dfY, batch_size=32, epochs=epochs, validation_data=(val_dfX, val_dfY)) graf_model(train_history_tam2) precision(model2 )
Titanic - Machine Learning from Disaster
4,005,064
test_id = test['id'] columns = {'id', 'location'} train = train.drop(columns = columns) test = test.drop(columns = columns) train['keyword'] = train['keyword'].fillna('unknown') test['keyword'] = test['keyword'].fillna('unknown') train['text'] = train['text'] + ' ' + train['keyword'] test['text'] = test['text'] + ' ' + test['keyword'] columns = {'keyword'} train = train.drop(columns = columns) test = test.drop(columns = columns) total = train.append(test )<feature_engineering>
arqFinal = [1024, 1024, 1024] modelF = None modelF = func_model(arqFinal) print(modelF.summary()) train_history_tamF = modelF.fit(train_dfX, train_dfY, batch_size=32, epochs=epochs, validation_data=(val_dfX, val_dfY)) graf_model(train_history_tamF) precision(modelF, True )
Titanic - Machine Learning from Disaster
4,005,064
total['unique word count'] = total['text'].apply(lambda x: len(set(x.split()))) total['stopword count'] = total['text'].apply(lambda x: len([i for i in x.lower().split() if i in wordcloud.STOPWORDS])) total['stopword ratio'] = total['stopword count'] / total['word count'] total['punctuation count'] = total['text'].apply(lambda x: len([i for i in str(x)if i in string.punctuation])) train = total[:len(train)] disaster = train['target'] == 1 fig, axes = plt.subplots(4, figsize=(20, 30)) graph1 = sns.kdeplot(train.loc[~disaster]['unique word count'], shade = True, label = 'Not Disaster', ax=axes[0]) graph1 = sns.kdeplot(train.loc[disaster]['unique word count'], shade = True, label = 'Disaster', ax=axes[0]) graph1.set_title('Distribution of Unique Word Count') graph2 = sns.kdeplot(train.loc[~disaster]['stopword count'], shade = True, label = 'Not Disaster', ax=axes[1]) graph2 = sns.kdeplot(train.loc[disaster]['stopword count'], shade = True, label = 'Disaster', ax=axes[1]) graph2.set_title('Distribution of Stopword Word Count') graph3 = sns.kdeplot(train.loc[~disaster]['punctuation count'], shade = True, label = 'Not Disaster', ax=axes[2], bw = 1) graph3 = sns.kdeplot(train.loc[disaster]['punctuation count'], shade = True, label = 'Disaster', ax=axes[2], bw = 1) graph3.set_title('Distribution of Punctuation Count') graph4 = sns.kdeplot(train.loc[~disaster]['stopword ratio'], shade = True, label = 'Not Disaster', ax=axes[3], bw =.05) graph4 = sns.kdeplot(train.loc[disaster]['stopword ratio'], shade = True, label = 'Disaster', ax=axes[3], bw =.05) graph4.set_title('Distribution of Stopword Ratio') fig.tight_layout() plt.show()<string_transform>
def func_model_reg() : np.random.seed(42) random_seed = 42 inp = Input(shape=(train_dfX.shape[1],)) x=Dropout(0.1 )(inp) x=Dense(1024, activation="relu", kernel_initializer=initializers.RandomNormal(seed=random_seed), bias_initializer='zeros', kernel_regularizer=regularizers.l2(0.01))(x) x=Dropout(0.7 )(x) x=Dense(1024, activation="relu", kernel_initializer=initializers.RandomNormal(seed=random_seed), bias_initializer='zeros', kernel_regularizer=regularizers.l2(0.01))(x) x=Dropout(0.7 )(x) x=Dense(512, activation="relu", kernel_initializer=initializers.RandomNormal(seed=random_seed), bias_initializer='zeros', kernel_regularizer=regularizers.l2(0.01))(x) x=Dropout(0.9 )(x) x=Dense(1, activation="sigmoid", kernel_initializer=initializers.RandomNormal(seed=random_seed), bias_initializer='zeros' )(x) model = Model(inputs=inp, outputs=x) model.compile(loss='binary_crossentropy', optimizer=optimizers.Adam(lr=0.0002), metrics=['binary_accuracy']) return model
Titanic - Machine Learning from Disaster
4,005,064
def remove_punctuation(x): return x.translate(str.maketrans('', '', string.punctuation)) def remove_stopwords(x): return ' '.join([i for i in x.split() if i not in wordcloud.STOPWORDS]) def remove_less_than(x): return ' '.join([i for i in x.split() if len(i)> 3]) def remove_non_alphabet(x): return ' '.join([i for i in x.split() if i.isalpha() ]) def strip_all_entities(x): return ' '.join(re.sub("(@[A-Za-z0-9]+)|([^0-9A-Za-z \t])|(\w+:\/\/\S+)"," ",x ).split() )<drop_column>
modelReg = None modelReg = func_model_reg() train_history_tamReg = modelReg.fit(train_dfX, train_dfY, batch_size=32, epochs=epochs, validation_data=(val_dfX, val_dfY)) graf_model(train_history_tamReg) precision(modelReg )
Titanic - Machine Learning from Disaster
4,005,064
<string_transform><EOS>
y_test = modelReg.predict(test_df) submission['Survived'] = y_test.round().astype(int) submission.to_csv('submission.csv', index=False )
Titanic - Machine Learning from Disaster
6,488,543
<SOS> metric: categorizationaccuracy Kaggle data source: titanic-machine-learning-from-disaster<feature_engineering>
%matplotlib inline py.init_notebook_mode(connected=True) warnings.filterwarnings('ignore') GradientBoostingClassifier, ExtraTreesClassifier)
Titanic - Machine Learning from Disaster