kernel_id int64 24.2k 23.3M | prompt stringlengths 8 1.85M | completetion stringlengths 1 182k | comp_name stringlengths 5 57 |
|---|---|---|---|
1,249,981 | x_train.fillna(0,inplace=True)
x_test.fillna(0,inplace=True )<import_modules> | out_file = 'random_search_trials.csv'
of_connection = open(out_file, 'w')
writer = csv.writer(of_connection)
headers = ['score', 'hyperparameters', 'iteration']
writer.writerow(headers)
of_connection.close() | Home Credit Default Risk |
1,249,981 | from sklearn.linear_model import LinearRegression
from sklearn.preprocessing import StandardScaler<prepare_x_and_y> | def random_search(param_grid, out_file, max_evals = MAX_EVALS):
results = pd.DataFrame(columns = ['score', 'params', 'iteration'],
index = list(range(MAX_EVALS)))
for i in range(MAX_EVALS):
random_params = {k: random.sample(v, 1)[0] for k, v in param_grid.items() }
random_params['subsample'] = 1.0 if random_params['... | Home Credit Default Risk |
1,249,981 | x_scaler = StandardScaler()
y_scaler = StandardScaler()<normalization> | def grid_search(param_grid, out_file, max_evals = MAX_EVALS):
results = pd.DataFrame(columns = ['score', 'params', 'iteration'],
index = list(range(MAX_EVALS)))
keys, values = zip(*param_grid.items())
i = 0
for v in itertools.product(*values):
parameters = dict(zip(keys, v))
parameters['subsample'] = 1.0 if paramet... | Home Credit Default Risk |
1,249,981 | x_train_scaler = x_scaler.fit_transform(x_train)
x_test_scaler = x_scaler.transform(x_test )<choose_model_class> | random_results = pd.read_csv('.. /input/home-credit-model-tuning/random_search_trials_1000.csv')
grid_results = pd.read_csv('.. /input/home-credit-model-tuning/grid_search_trials_1000.csv' ) | Home Credit Default Risk |
1,249,981 | lr = LinearRegression()<train_model> | grid_results['hyperparameters'] = grid_results['hyperparameters'].map(ast.literal_eval)
random_results['hyperparameters'] = random_results['hyperparameters'].map(ast.literal_eval ) | Home Credit Default Risk |
1,249,981 | lr.fit(x_train_scaler,y_train )<predict_on_test> | def evaluate(results, name):
results = results.sort_values('score', ascending = False ).reset_index(drop = True)
print('The highest cross validation score from {} was {:.5f} found on iteration {}.'.format(name, results.loc[0, 'score'], results.loc[0, 'iteration']))
hyperparameters = results.loc[0, 'hyperparameters']... | Home Credit Default Risk |
1,249,981 | pred = lr.predict(x_test_scaler )<create_dataframe> | grid_hyp = evaluate(grid_results, name = 'grid search' ) | Home Credit Default Risk |
1,249,981 | sub = pd.DataFrame(data = {'Id' : test.Id, 'SalePrice' :pred} )<save_to_csv> | random_hyp = evaluate(random_results, name = 'random search' ) | Home Credit Default Risk |
1,249,981 | sub.to_csv('submission.csv',index = False )<load_from_csv> | alt.renderers.enable('notebook' ) | Home Credit Default Risk |
1,249,981 | train = pd.read_csv('.. /input/train.csv')
test = pd.read_csv('.. /input/test.csv' )<import_modules> | random_hyp['search'] = 'random'
grid_hyp['search'] = 'grid'
hyp = random_hyp.append(grid_hyp)
hyp.head() | Home Credit Default Risk |
1,249,981 | from sklearn.linear_model import LinearRegression
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import PolynomialFeatures<drop_column> | best_grid_hyp = grid_hyp.iloc[grid_hyp['score'].idxmax() ].copy()
best_random_hyp = random_hyp.iloc[random_hyp['score'].idxmax() ].copy() | Home Credit Default Risk |
1,249,981 | pipe = Pipeline([
('scaler', StandardScaler()),
('poly', PolynomialFeatures()),
('model', LinearRegression()),
] )<prepare_x_and_y> | print('Average validation score of grid search = {:.5f}.'.format(np.mean(grid_hyp['score'])))
print('Average validation score of random search = {:.5f}.'.format(np.mean(random_hyp['score'])) ) | Home Credit Default Risk |
1,249,981 | X_train = train[['LotFrontage','LotArea']].fillna(0)
X_test = test[['LotFrontage','LotArea']].fillna(0)
y_train = train['SalePrice']<import_modules> | random_hyp['score'] = random_hyp['score'].astype(float)
best_random_hyp = random_hyp.loc[0, :].copy() | Home Credit Default Risk |
1,249,981 | from sklearn.model_selection import cross_val_score
from sklearn.metrics import mean_squared_error<compute_train_metric> | train = pd.read_csv('.. /input/home-credit-simple-featuers/simple_features_train.csv')
test = pd.read_csv('.. /input/home-credit-simple-featuers/simple_features_test.csv')
test_ids = test['SK_ID_CURR']
train_labels = np.array(train['TARGET'].astype(np.int32)).reshape(( -1,))
train = train.drop(columns = ['SK_ID_CURR'... | Home Credit Default Risk |
1,249,981 | scores = cross_val_score(pipe, X_train, y_train, scoring='neg_mean_squared_error', cv=10 )<feature_engineering> | train_set = lgb.Dataset(train, label = train_labels)
hyperparameters = dict(**random_results.loc[0, 'hyperparameters'])
del hyperparameters['n_estimators']
cv_results = lgb.cv(hyperparameters, train_set,
num_boost_round = 10000, early_stopping_rounds = 100,
metrics = 'auc', nfold = N_FOLDS ) | Home Credit Default Risk |
1,249,981 | np.mean(pow(-scores, 0.5))<train_model> | print('The cross validation score on the full dataset = {:.5f} with std: {:.5f}.'.format(
cv_results['auc-mean'][-1], cv_results['auc-stdv'][-1]))
print('Number of estimators = {}.'.format(len(cv_results['auc-mean'])) ) | Home Credit Default Risk |
1,249,981 | pipe.fit(X_train, y_train )<predict_on_test> | model = lgb.LGBMClassifier(n_estimators = len(cv_results['auc-mean']), **hyperparameters)
model.fit(train, train_labels)
preds = model.predict_proba(test)[:, 1] | Home Credit Default Risk |
1,249,981 | preds = pipe.predict(X_test )<save_to_csv> | submission = pd.DataFrame({'SK_ID_CURR': test_ids, 'TARGET': preds})
submission.to_csv('submission_simple_features_random.csv', index = False ) | Home Credit Default Risk |
1,254,160 | sub = pd.DataFrame({'Id': test.Id, 'SalePrice': preds})
sub.to_csv('submission.csv', index=False )<set_options> | %matplotlib inline | Home Credit Default Risk |
1,254,160 | warnings.filterwarnings('ignore')
<load_from_csv> | from sklearn.model_selection import GridSearchCV
from skopt import BayesSearchCV | Home Credit Default Risk |
1,254,160 | data_path = '.. /input/kaggledays-paris/'
train = pd.read_csv(data_path + 'train.csv',encoding='UTF-8')
test = pd.read_csv(data_path + 'test.csv',encoding='UTF-8')
colorfile = pd.read_csv('.. /input/color-feature/color_transcode2.csv',encoding='UTF-8')
train = train.merge(colorfile, on='color', how='left')
test = t... | import lightgbm as lgb | Home Credit Default Risk |
1,254,160 | print('Preprocessing text...')
tfidf = TfidfVectorizer(max_features=5,norm='l2',)
tfidf.fit(df_all[ 'en_US_description'].astype(str ).values)
tfidf_all = np.array(tfidf.transform(df_all[ 'en_US_description'].astype(str ).values ).toarray() , dtype=np.float16)
for i in tqdm(range(5)) :
df_all['en_US_description_tfid... | import gc | Home Credit Default Risk |
1,254,160 | print('Label Encoder...')
cols = [f_ for f_ in df_all.columns if df_all[f_].dtype == 'object' and 'sku' not in f_ and 'ID' not in f_]
print(cols)
cnt = 0
for c in tqdm(cols):
le = LabelEncoder()
df_all[c] = le.fit_transform(df_all[c].astype(str))
cnt += 1
del le
print('len(cols)= {}'.format(cnt))
train = df_all.loc[d... | from sklearn.model_selection import train_test_split, KFold, StratifiedKFold
from sklearn.metrics import accuracy_score, confusion_matrix
from sklearn.feature_selection import VarianceThreshold
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder, MinMaxScaler | Home Credit Default Risk |
1,254,160 | vimages = pd.read_csv(data_path + 'vimages.csv')
pca = PCA(n_components=10)
vpca = pca.fit_transform(vimages.drop('sku_hash', axis=1 ).values)
vimages_pca = vimages[['sku_hash']]
for i in tqdm(range(vpca.shape[1])) :
vimages_pca['dim_pca_{}'.format(i)] = vpca[:, i]
train = train.merge(vimages_pca, on='sku_hash', how... | def status_print(optim_result):
all_models = pd.DataFrame(bayes_cv_tuner.cv_results_)
best_params = pd.Series(bayes_cv_tuner.best_params_)
print('Model
Best ROC-AUC: {}
Best params: {}
'.format(
len(all_models),
np.round(bayes_cv_tuner.best_score_, 4),
bayes_cv_tuner.best_params_
))
clf_name = bayes_cv_tuner.esti... | Home Credit Default Risk |
1,254,160 | sales = pd.read_csv(data_path + 'sales.csv')
sales = sales.merge(df_all[['sku_hash','model','function']], on='sku_hash', how='left')
sales = sales.sort_values('Date', ascending=True)
salesg = sales.groupby('sku_hash' ).agg({
'sales_quantity': ['sum', 'mean'],
'Month_transaction': ['last'],
} ).reset_index()
salesg.c... | data = pd.read_csv(".. /input/application_train.csv")
test = pd.read_csv('.. /input/application_test.csv')
prev = pd.read_csv('.. /input/previous_application.csv')
buro = pd.read_csv('.. /input/bureau.csv')
buro_balance = pd.read_csv('.. /input/bureau_balance.csv')
credit_card = pd.read_csv('.. /input/credit_card_... | Home Credit Default Risk |
1,254,160 | agg_tmp = sales.groupby(['model','Date'] ).agg(aggs)
agg_tmp.columns = [ '{}_{}'.format(k, agg)for k in aggs.keys() for agg in aggs[k]]
agg_tmp=agg_tmp.reset_index()
agg_tmp = agg_tmp.sort_values('Date', ascending=True)
XX_sales_BYMODEL = agg_tmp.groupby(['model'])['sales_quantity_sum'].apply(lambda x: sales_features... | y = data["TARGET"]
del data["TARGET"] | Home Credit Default Risk |
1,254,160 | def extract_features(df):
df['month_real'] =(df['month']+df['Month_transaction_last'] ).apply(lambda x: x%12)
pass
print('Extracting month_real features for train:')
extract_features(train)
print('Extracting month_real features for test:')
extract_features(test)
print('train ',train.shape)
print('test ',test.shap... | categorical_feature = [col for col in data.columns if data[col].dtype == 'object'] | Home Credit Default Risk |
1,254,160 | y = np.log1p(train['target'].values)
f = [
'en_US_description_tfidf_0', 'en_US_description_tfidf_1', 'en_US_description_tfidf_2', 'en_US_description_tfidf_3', 'en_US_description_tfidf_4',
'dim_pca_0', 'dim_pca_1', 'dim_pca_2', 'dim_pca_3','dim_pca_4',
'dim_pca_5', 'dim_pca_6', 'dim_pca_7', 'dim_pca_8','dim_pca_9',
'sa... | one_hot_df = pd.concat([data, test])
one_hot_df = pd.get_dummies(one_hot_df, columns=categorical_feature ) | Home Credit Default Risk |
1,254,160 | clfs, importances, err, oof_predict = train_classifiers(train,y,f)
<save_to_csv> | data = one_hot_df.iloc[:data.shape[0],:]
test = one_hot_df.iloc[data.shape[0]:,] | Home Credit Default Risk |
1,254,160 | preds_ = None
for clf in clfs:
if preds_ is None:
preds_ = clf.predict(test[f], num_iteration=clf.best_iteration_)
else:
preds_ += clf.predict(test[f], num_iteration=clf.best_iteration_)
preds_ = preds_ / len(clfs)
preds_ = np.expm1(preds_)
subm = pd.DataFrame()
subm['ID'] = test['ID'].values
subm['target'] = preds... | buro_grouped_size = buro_balance.groupby("SK_ID_BUREAU")["MONTHS_BALANCE"].size()
buro_grouped_max = buro_balance.groupby("SK_ID_BUREAU")["MONTHS_BALANCE"].max()
buro_grouped_min = buro_balance.groupby("SK_ID_BUREAU")["MONTHS_BALANCE"].min() | Home Credit Default Risk |
1,254,160 | import torch
import torch.nn as nn
import pandas as pd
import numpy as np
import xgboost as xgb
from sklearn import preprocessing
from sklearn.model_selection import train_test_split
from sklearn.model_selection import TimeSeriesSplit,KFold, StratifiedKFold
from time import time
import pickle
import datetime as dt
from... | buro_counts = buro_balance.groupby("SK_ID_BUREAU")["STATUS"].value_counts(normalize=False)
buro_counts_unstacked = buro_counts.unstack("STATUS")
| Home Credit Default Risk |
1,254,160 | import torch
import torch.nn as nn
import torch.optim as optim
from torch.utils.data import Dataset, DataLoader, random_split<load_from_csv> | buro_counts_unstacked.columns = ['STATUS_0', 'STATUS_1','STATUS_2','STATUS_3','STATUS_4','STATUS_5','STATUS_C','STATUS_X',]
buro_counts_unstacked["MONTHS_COUNT"] = buro_grouped_size
buro_counts_unstacked["MONTHS_MIN"] = buro_grouped_min
buro_counts_unstacked["MONTHS_MAX"] = buro_grouped_max | Home Credit Default Risk |
1,254,160 | def init_read_big(train_path, test_path, train_big_path):
train = pd.read_csv(train_path)
test = pd.read_csv(test_path)
train_big = pd.read_csv(train_big_path)
return train, test, train_big
def init_read_small(train_path, test_path):
train = pd.read_csv(train_path)
test = pd.read_csv(test_path)
return train, test
... | buro = buro.join(buro_counts_unstacked, how='left', on='SK_ID_BUREAU' ) | Home Credit Default Risk |
1,254,160 | train_path = '/kaggle/input/2019s-uts-data-analytics-assignment-3/Assignment3_TrainingSet.csv'
test_path = '/kaggle/input/2019s-uts-data-analytics-assignment-3/Assignment3_TestSet.csv'
train, test = init_read_small(train_path, test_path)
train, test, y = data_preprocessing(train, test )<normalization> | prev_cat_features = [pcol for pcol in prev.columns if prev[pcol].dtype == "object"]
prev = pd.get_dummies(prev, columns=prev_cat_features)
| Home Credit Default Risk |
1,254,160 | sc = StandardScaler()
train = sc.fit_transform(train)
test = sc.fit_transform(test )<train_model> | avg_prev = prev.groupby("SK_ID_CURR" ).mean()
| Home Credit Default Risk |
1,254,160 |
<normalization> | cnt_prev = prev[["SK_ID_CURR", "SK_ID_PREV"]].groupby("SK_ID_CURR" ).count()
avg_prev["nb_app"] = cnt_prev["SK_ID_PREV"]
del avg_prev["SK_ID_PREV"] | Home Credit Default Risk |
1,254,160 | class Net(nn.Module):
def __init__(self, D_in, H=15, D_out=1):
super().__init__()
self.fc1 = nn.Linear(D_in, H)
self.fc2 = nn.Linear(H, D_out)
self.relu = nn.ReLU()
def forward(self, x):
x = self.fc1(x)
x = self.relu(x)
x = self.fc2(x)
return x.squeeze()<find_best_params> | buro_cat_features = [bcol for bcol in buro.columns if buro[bcol].dtype == 'object']
buro = pd.get_dummies(buro, columns=buro_cat_features)
avg_buro = buro.groupby('SK_ID_CURR' ).mean()
avg_buro['buro_count'] = buro[['SK_ID_BUREAU', 'SK_ID_CURR']].groupby('SK_ID_CURR' ).count() ['SK_ID_BUREAU']
del avg_buro['SK_ID_BURE... | Home Credit Default Risk |
1,254,160 | class EarlyStopping(object):
def __init__(self, mode='min', min_delta=0, patience=10, percentage=False):
self.mode = mode
self.min_delta = min_delta
self.patience = patience
self.best = None
self.num_bad_epochs = 0
self.is_better = None
self._init_is_better(mode, min_delta, percentage)
if patience == 0:
self.is_better... | le = LabelEncoder()
POS_CASH['NAME_CONTRACT_STATUS'] = le.fit_transform(POS_CASH['NAME_CONTRACT_STATUS'].astype(str))
nunique_status = POS_CASH[['SK_ID_CURR', 'NAME_CONTRACT_STATUS']].groupby('SK_ID_CURR' ).nunique()
nunique_status2 = POS_CASH[['SK_ID_CURR', 'NAME_CONTRACT_STATUS']].groupby('SK_ID_CURR' ).max()
nunique... | Home Credit Default Risk |
1,254,160 | device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu" )<split> | POS_CASH['NUNIQUE_STATUS'] = nunique_status['NAME_CONTRACT_STATUS']
POS_CASH['NUNIQUE_STATUS2'] = nunique_status2['NAME_CONTRACT_STATUS']
POS_CASH.drop(['SK_ID_PREV', 'NAME_CONTRACT_STATUS'], axis=1, inplace=True ) | Home Credit Default Risk |
1,254,160 | n_fold = 5
random_state = 999
D_in, H = 30, 15
models = []
train_no = 1
training_cycle = 20000
repetition = 1
training_sequence = 1
for A in range(repetition):
kf = StratifiedKFold(n_splits = n_fold , shuffle = True, random_state = random_state+A)
for train_index, val_index in kf.split(train, y):
train_X = train[train... | le =LabelEncoder()
credit_card['NAME_CONTRACT_STATUS'] = le.fit_transform(credit_card['NAME_CONTRACT_STATUS'].astype(str))
nunique_status = credit_card[['SK_ID_CURR', 'NAME_CONTRACT_STATUS']].groupby('SK_ID_CURR' ).nunique()
nunique_status2 = credit_card[['SK_ID_CURR', 'NAME_CONTRACT_STATUS']].groupby('SK_ID_CURR' ).ma... | Home Credit Default Risk |
1,254,160 |
<define_variables> | avg_payments = payments.groupby("SK_ID_CURR" ).mean()
avg_payments2 = payments.groupby("SK_ID_CURR" ).max()
avg_payments3 = payments.groupby("SK_ID_CURR" ).min() | Home Credit Default Risk |
1,254,160 | test_torch = torch.tensor(test)
pred_df = sum([net(test_torch.float())for net in models])/(n_fold*repetition )<predict_on_test> | del avg_payments["SK_ID_PREV"] | Home Credit Default Risk |
1,254,160 | test_predict_numpy = pred_df.detach().numpy()<feature_engineering> | data = data.merge(right=avg_prev.reset_index() , how="left", on="SK_ID_CURR")
test = test.merge(right=avg_prev.reset_index() , how="left", on="SK_ID_CURR")
| Home Credit Default Risk |
1,254,160 | for A in range(len(test_predict_numpy)) :
if test_predict_numpy[A] > np.mean(test_predict_numpy):
test_predict_numpy[A] = 1
else:
test_predict_numpy[A] = 0<count_values> | data = data.merge(right=avg_buro.reset_index() , how='left', on='SK_ID_CURR')
test = test.merge(right=avg_buro.reset_index() , how='left', on='SK_ID_CURR' ) | Home Credit Default Risk |
1,254,160 | sum(test_predict_numpy )<save_to_csv> | data = data.merge(right=POS_CASH.groupby("SK_ID_CURR" ).mean().reset_index() , how="left", on="SK_ID_CURR")
test = test.merge(right=POS_CASH.groupby("SK_ID_CURR" ).mean().reset_index() , how="left", on="SK_ID_CURR" ) | Home Credit Default Risk |
1,254,160 | sample = pd.read_csv('/kaggle/input/2019s-uts-data-analytics-assignment-3/Assignment3_Random_Submission-Kaggle.csv')
sample.QuoteConversion_Flag = test_predict_numpy
sample.to_csv('pytorch_model.csv', index=False)
print("<load_from_csv> | data = data.merge(credit_card.groupby('SK_ID_CURR' ).mean().reset_index() , how='left', on='SK_ID_CURR')
test = test.merge(credit_card.groupby('SK_ID_CURR' ).mean().reset_index() , how='left', on='SK_ID_CURR')
| Home Credit Default Risk |
1,254,160 | train=pd.read_csv('/kaggle/input/ec524-heart-disease/train.csv')
test=pd.read_csv('/kaggle/input/ec524-heart-disease/test.csv' )<count_missing_values> | data = data.merge(right=avg_payments.reset_index() , how='left', on='SK_ID_CURR')
test = test.merge(right=avg_payments.reset_index() , how='left', on='SK_ID_CURR')
| Home Credit Default Risk |
1,254,160 | train.isna().sum()<data_type_conversions> | data = data.merge(right=avg_payments2.reset_index() , how='left', on='SK_ID_CURR')
test = test.merge(right=avg_payments2.reset_index() , how='left', on='SK_ID_CURR' ) | Home Credit Default Risk |
1,254,160 | train.fillna(train.mean() ,inplace=True )<count_missing_values> | data = data.merge(right=avg_payments3.reset_index() , how='left', on='SK_ID_CURR')
test = test.merge(right=avg_payments3.reset_index() , how='left', on='SK_ID_CURR' ) | Home Credit Default Risk |
1,254,160 | train.isna().sum()<count_missing_values> | test = test[test.columns[data.isnull().mean() < 0.85]]
data = data[data.columns[data.isnull().mean() < 0.85]] | Home Credit Default Risk |
1,254,160 | test.isna().sum()<drop_column> | excluded_feats = ["SK_ID_CURR"]
features = [f for f in data.columns if f not in excluded_feats] | Home Credit Default Risk |
1,254,160 | train=train.drop(['id'],axis=1 )<prepare_x_and_y> | Home Credit Default Risk | |
1,254,160 | x=train.drop(['heart_disease'],axis=1)
y=train['heart_disease'].values<normalization> | lgbm_params = {
"boosting":"dart",
"application":"binary",
"learning_rate": 0.1,
'reg_alpha':0.01,
'reg_lambda': 0.01,
"n_estimators":10000,
"max_depth":7,
"num_leaves":100,
"max_bin":225,
"drop_rate":0.02
} | Home Credit Default Risk |
1,254,160 | scaler=StandardScaler()
x=scaler.fit_transform(x)
<choose_model_class> | model = lgb.LGBMClassifier(application="binary", boosting_type=lgbm_params["boosting"],
learning_rate=lgbm_params["learning_rate"],n_estimators=lgbm_params["n_estimators"],drop_rate=lgbm_params["drop_rate"],
num_leaves=lgbm_params["num_leaves"], max_depth=lgbm_params["max_depth"],
max_bin=lgbm_params["max_bin"])
| Home Credit Default Risk |
1,254,160 | lr=LogisticRegression()
rf=RandomForestClassifier()
vote=VotingClassifier(estimators=[('lr',lr),('rf',rf)] )<train_model> | feature_importances = np.zeros(data.shape[1])
for i in range(2):
train_data, test_data, train_y, test_y = train_test_split(data, y, test_size=0.2, random_state=i)
model.fit(train_data, train_y, early_stopping_rounds=100, eval_set=[(test_data, test_y)], eval_metric='auc', verbose=200)
feature_importances += model.fea... | Home Credit Default Risk |
1,254,160 | vote.fit(x,y )<compute_test_metric> | threshold = 0.98
features_to_keep = list(norm_feature_importances[norm_feature_importances['cumulative_importance'] < threshold]['feature'])
data = data[features_to_keep]
test = test[features_to_keep] | Home Credit Default Risk |
1,254,160 | vote.score(x,y )<drop_column> | n_folds = 5
k_fold = KFold(n_splits=n_folds, shuffle=False, random_state=50)
feature_importances_values = np.zeros(data.shape[1])
test_predictions = np.zeros(test.shape[0])
out_of_fold = np.zeros(data.shape[0])
valid_scores = []
train_scores = []
for train_indices, test_indices in k_fold.split(data):
train_data, tr... | Home Credit Default Risk |
1,254,160 | test_=test.drop(['id'],axis=1 )<normalization> | submission = pd.DataFrame({'SK_ID_CURR': test["SK_ID_CURR"], 'TARGET': test_predictions})
submission.to_csv("submissions.csv", index=False ) | Home Credit Default Risk |
1,233,526 | test_scaled=scaler.fit_transform(test_ )<predict_on_test> | import gc | Home Credit Default Risk |
1,233,526 | y_pred=vote.predict(test_scaled )<create_dataframe> | from sklearn.model_selection import train_test_split, KFold
from sklearn.metrics import accuracy_score, confusion_matrix
from sklearn.feature_selection import VarianceThreshold
from sklearn.preprocessing import LabelEncoder, MinMaxScaler | Home Credit Default Risk |
1,233,526 | sub=pd.DataFrame()
sub['id']=test['id']
sub['heart_disease']=y_pred
<save_to_csv> | data = pd.read_csv(".. /input/application_train.csv")
test = pd.read_csv('.. /input/application_test.csv')
prev = pd.read_csv('.. /input/previous_application.csv')
buro = pd.read_csv('.. /input/bureau.csv')
buro_balance = pd.read_csv('.. /input/bureau_balance.csv')
credit_card = pd.read_csv('.. /input/credit_card_... | Home Credit Default Risk |
1,233,526 | sub.to_csv('sub.csv',index=False )<set_options> | y = data["TARGET"]
del data["TARGET"] | Home Credit Default Risk |
1,233,526 | img_size =(224, 224)
img_array_list = []
cls_list = []
dir = '.. /input/train/PNEUMONIA'
img_list = glob.glob(dir + '/*.jpeg')
for i in img_list:
img = load_img(i, color_mode='grayscale', target_size=(img_size))
img_array = img_to_array(img)/ 255
img_array_list.append(img_array)
cls_list.append(1)
dir = '.. /input/... | categorical_feature = [col for col in data.columns if data[col].dtype == 'object'] | Home Credit Default Risk |
1,233,526 | X_learn, X_valid, y_learn, y_valid = train_test_split(X_train, y_train, test_size=0.3, random_state=0)
shape = X_learn.shape
X_learn = np.reshape(X_learn,(shape[0], shape[1] * shape[2] * shape[3]))
ros = RandomOverSampler(random_state=0)
X_res, y_res = ros.fit_resample(X_learn, y_learn)
X_res = np.reshape(X_res,(X_r... | one_hot_df = pd.concat([data, test])
one_hot_df = pd.get_dummies(one_hot_df, columns=categorical_feature ) | Home Credit Default Risk |
1,233,526 | def auc(y_true, y_pred):
auc = tf.py_func(roc_auc_score,(y_true, y_pred), tf.double)
return auc
model = Sequential()
model.add(Conv2D(filters=64, kernel_size=(3, 3), activation='relu', input_shape=(img_size[0], img_size[1], 1)))
model.add(Conv2D(filters=64, kernel_size=(3, 3), activation='relu'))
model.add(MaxPooling... | data = one_hot_df.iloc[:data.shape[0],:]
test = one_hot_df.iloc[data.shape[0]:,] | Home Credit Default Risk |
1,233,526 | model.fit(X_res, Y_res, batch_size=32, epochs=10)
model.evaluate(X_valid, Y_valid, batch_size=32 )<train_model> | buro_grouped_size = buro_balance.groupby("SK_ID_BUREAU")["MONTHS_BALANCE"].size()
buro_grouped_max = buro_balance.groupby("SK_ID_BUREAU")["MONTHS_BALANCE"].max()
buro_grouped_min = buro_balance.groupby("SK_ID_BUREAU")["MONTHS_BALANCE"].min() | Home Credit Default Risk |
1,233,526 | shape = X_train.shape
X_train = np.reshape(X_train,(shape[0], shape[1] * shape[2] * shape[3]))
ros = RandomOverSampler(random_state=0)
X_res, y_res = ros.fit_resample(X_train, y_train)
X_train = np.reshape(X_train,(shape[0], shape[1], shape[2], shape[3]))
X_res = np.reshape(X_res,(X_res.shape[0], shape[1], shape[2], ... | buro_counts = buro_balance.groupby("SK_ID_BUREAU")["STATUS"].value_counts(normalize=False)
buro_counts_unstacked = buro_counts.unstack("STATUS")
| Home Credit Default Risk |
1,233,526 | model.fit(X_res, Y_res, batch_size=32, epochs=10 )<normalization> | buro_counts_unstacked.columns = ['STATUS_0', 'STATUS_1','STATUS_2','STATUS_3','STATUS_4','STATUS_5','STATUS_C','STATUS_X',]
buro_counts_unstacked["MONTHS_COUNT"] = buro_grouped_size
buro_counts_unstacked["MONTHS_MIN"] = buro_grouped_min
buro_counts_unstacked["MONTHS_MAX"] = buro_grouped_max | Home Credit Default Risk |
1,233,526 | img_array_list = []
dir = '.. /input/test'
img_list = glob.glob(dir + '/*.jpeg')
img_list.sort()
for i in img_list:
img = load_img(i, color_mode='grayscale', target_size=(img_size))
img_array = img_to_array(img)/ 255
img_array_list.append(img_array)
X_test = np.array(img_array_list)
print(X_test.shape )<predict_on_t... | buro = buro.join(buro_counts_unstacked, how='left', on='SK_ID_BUREAU' ) | Home Credit Default Risk |
1,233,526 | predict = model.predict(X_test)[:, 1]<save_to_csv> | prev_cat_features = [pcol for pcol in prev.columns if prev[pcol].dtype == "object"]
prev = pd.get_dummies(prev, columns=prev_cat_features)
prev.head(10 ) | Home Credit Default Risk |
1,233,526 | submit = pd.read_csv('.. /input/sampleSubmission.csv')
submit['pneumonia'] = predict
submit.to_csv('submission.csv', index=False )<import_modules> | avg_prev = prev.groupby("SK_ID_CURR" ).mean()
avg_prev.head(10 ) | Home Credit Default Risk |
1,233,526 | import numpy as np
from sklearn.metrics.pairwise import cosine_similarity
from sklearn.feature_extraction.text import TfidfVectorizer<define_variables> | cnt_prev = prev[["SK_ID_CURR", "SK_ID_PREV"]].groupby("SK_ID_CURR" ).count()
avg_prev["nb_app"] = cnt_prev["SK_ID_PREV"]
del avg_prev["SK_ID_PREV"] | Home Credit Default Risk |
1,233,526 | root_path = '.. /input/homework4_v2/'
max_df = 0.95
min_df = 5
smooth_idf = True
sublinear_tf = True
alpha = 1
beta = 0.75
gamma = 0.15
rel_count = 5
nrel_count = 1
iters = 5<define_variables> | buro_cat_features = [bcol for bcol in buro.columns if buro[bcol].dtype == 'object']
buro = pd.get_dummies(buro, columns=buro_cat_features)
avg_buro = buro.groupby('SK_ID_CURR' ).mean()
avg_buro['buro_count'] = buro[['SK_ID_BUREAU', 'SK_ID_CURR']].groupby('SK_ID_CURR' ).count() ['SK_ID_BUREAU']
del avg_buro['SK_ID_BURE... | Home Credit Default Risk |
1,233,526 | with open(root_path + 'doc_list.txt')as file:
doc_list = [line.rstrip() for line in file]
with open(root_path + 'query_list.txt')as file:
query_list = [line.rstrip() for line in file]<string_transform> | le = LabelEncoder()
POS_CASH['NAME_CONTRACT_STATUS'] = le.fit_transform(POS_CASH['NAME_CONTRACT_STATUS'].astype(str))
nunique_status = POS_CASH[['SK_ID_CURR', 'NAME_CONTRACT_STATUS']].groupby('SK_ID_CURR' ).nunique()
nunique_status2 = POS_CASH[['SK_ID_CURR', 'NAME_CONTRACT_STATUS']].groupby('SK_ID_CURR' ).max()
nunique... | Home Credit Default Risk |
1,233,526 | documents, queries = [], []
for doc_name in doc_list:
with open(root_path + 'Document/' + doc_name)as file:
doc = ' '.join([word for line in file.readlines() [3:] for word in line.split() [:-1]])
documents.append(doc)
for query_name in query_list:
with open(root_path + 'Query/' + query_name)as file:
query = ' '.join(... | POS_CASH['NUNIQUE_STATUS'] = nunique_status['NAME_CONTRACT_STATUS']
POS_CASH['NUNIQUE_STATUS2'] = nunique_status2['NAME_CONTRACT_STATUS']
POS_CASH.drop(['SK_ID_PREV', 'NAME_CONTRACT_STATUS'], axis=1, inplace=True ) | Home Credit Default Risk |
1,233,526 | vectorizer = TfidfVectorizer(max_df=max_df, min_df=min_df,
smooth_idf=smooth_idf, sublinear_tf=sublinear_tf)
doc_tfidfs = vectorizer.fit_transform(documents ).toarray()
query_vecs = vectorizer.transform(queries ).toarray()
cos_sim = cosine_similarity(query_vecs, doc_tfidfs)
rankings = np.flip(cos_sim.argsort() , axis... | le =LabelEncoder()
credit_card['NAME_CONTRACT_STATUS'] = le.fit_transform(credit_card['NAME_CONTRACT_STATUS'].astype(str))
nunique_status = credit_card[['SK_ID_CURR', 'NAME_CONTRACT_STATUS']].groupby('SK_ID_CURR' ).nunique()
nunique_status2 = credit_card[['SK_ID_CURR', 'NAME_CONTRACT_STATUS']].groupby('SK_ID_CURR' ).ma... | Home Credit Default Risk |
1,233,526 | for _ in range(iters):
rel_vecs = doc_tfidfs[rankings[:, :rel_count]].mean(axis=1)
nrel_vecs = doc_tfidfs[rankings[:, -nrel_count:]].mean(axis=1)
query_vecs = alpha * query_vecs + beta * rel_vecs - gamma * nrel_vecs
cos_sim = cosine_similarity(query_vecs, doc_tfidfs)
rankings = np.flip(cos_sim.argsort(axis=1), axis=... | avg_payments = payments.groupby("SK_ID_CURR" ).mean()
avg_payments2 = payments.groupby("SK_ID_CURR" ).max()
avg_payments3 = payments.groupby("SK_ID_CURR" ).min() | Home Credit Default Risk |
1,233,526 | with open('submission.csv', mode='w')as file:
file.write('Query,RetrievedDocuments
')
for query_name, ranking in zip(query_list, rankings):
ranked_docs = ' '.join([doc_list[idx] for idx in ranking])
file.write('%s,%s
' %(query_name, ranked_docs))<import_modules> | del avg_payments["SK_ID_PREV"] | Home Credit Default Risk |
1,233,526 | print(os.listdir(".. /input"))
<load_from_csv> | data = data.merge(right=avg_prev.reset_index() , how="left", on="SK_ID_CURR")
test = test.merge(right=avg_prev.reset_index() , how="left", on="SK_ID_CURR")
| Home Credit Default Risk |
1,233,526 | data = pd.read_csv(".. /input/train.csv")
test = pd.read_csv('.. /input/test.csv')
ids = pd.read_csv('.. /input/sampleSubmission.csv')
test.head(5 )<count_missing_values> | data = data.merge(right=avg_buro.reset_index() , how='left', on='SK_ID_CURR')
test = test.merge(right=avg_buro.reset_index() , how='left', on='SK_ID_CURR' ) | Home Credit Default Risk |
1,233,526 | data.isnull().any()<feature_engineering> | data = data.merge(right=POS_CASH.groupby("SK_ID_CURR" ).mean().reset_index() , how="left", on="SK_ID_CURR")
test = test.merge(right=POS_CASH.groupby("SK_ID_CURR" ).mean().reset_index() , how="left", on="SK_ID_CURR" ) | Home Credit Default Risk |
1,233,526 | data.loc[data['TotalCharges'].isna() ] = 0<count_missing_values> | data = data.merge(credit_card.groupby('SK_ID_CURR' ).mean().reset_index() , how='left', on='SK_ID_CURR')
test = test.merge(credit_card.groupby('SK_ID_CURR' ).mean().reset_index() , how='left', on='SK_ID_CURR')
| Home Credit Default Risk |
1,233,526 | test.isnull().any()<feature_engineering> | data = data.merge(right=avg_payments.reset_index() , how='left', on='SK_ID_CURR')
test = test.merge(right=avg_payments.reset_index() , how='left', on='SK_ID_CURR')
| Home Credit Default Risk |
1,233,526 | test.loc[test['TotalCharges'].isna() ] = 0<prepare_x_and_y> | data = data.merge(right=avg_payments2.reset_index() , how='left', on='SK_ID_CURR')
test = test.merge(right=avg_payments2.reset_index() , how='left', on='SK_ID_CURR' ) | Home Credit Default Risk |
1,233,526 | X = data.drop(['customerID','Churn'], 1)
x = pd.get_dummies(X)
x = x.drop(['gender_0','Partner_0','Dependents_0','PaymentMethod_0','MultipleLines_0','PaperlessBilling_0','StreamingMovies_0','PhoneService_0','InternetService_0','DeviceProtection_0','TechSupport_0','OnlineSecurity_0','OnlineBackup_0'],1)
y = pd.get_du... | data = data.merge(right=avg_payments3.reset_index() , how='left', on='SK_ID_CURR')
test = test.merge(right=avg_payments3.reset_index() , how='left', on='SK_ID_CURR' ) | Home Credit Default Risk |
1,233,526 | test_np = np.array(teste)
X=np.array(x)
Y=np.array(y )<split> | test = test[test.columns[data.isnull().mean() < 0.85]]
data = data[data.columns[data.isnull().mean() < 0.85]] | Home Credit Default Risk |
1,233,526 | x_Train,x_Test,y_Train,y_Test=train_test_split(X,Y,test_size=0.30 )<normalization> | excluded_feats = ["SK_ID_CURR"]
features = [f for f in data.columns if f not in excluded_feats] | Home Credit Default Risk |
1,233,526 | scaler = StandardScaler()
sdsfds = scaler.fit_transform(x_Train )<compute_train_metric> | folds = KFold(n_splits = 4, shuffle=True, random_state=546789)
oof_preds = np.zeros(data.shape[0])
sub_preds = np.zeros(test.shape[0] ) | Home Credit Default Risk |
1,233,526 | def score(model,x,y):
prob=model.predict_proba(x)
prob = prob[:, 1]
auc = roc_auc_score(y, prob)
print('AUC: {}
ROC_AUC: {}
{}'.format(model.score(x,y),auc,prob[:10]))<train_model> | for n_fold,(trn_idx, val_idx)in enumerate(folds.split(data)) :
trn_x, trn_y = data[features].iloc[trn_idx], y.iloc[trn_idx]
val_x, val_y = data[features].iloc[val_idx], y.iloc[val_idx]
clf = XGBClassifier(
objective="binary:logistic",
booster="gbtree",
eval_metric = "auc",
nthread = 4,
eta = 0.05,
gamma = 0,
max_depth... | Home Credit Default Risk |
1,233,526 | knn=KNeighborsClassifier(n_neighbors=3)
asa = knn.fit(x_Train,y_Train[:,1])
score(knn,x_Test,y_Test[:,1] )<predict_on_test> | print('Full AUC score %.6f' % roc_auc_score(y, oof_preds))
test['TARGET'] = sub_preds | Home Credit Default Risk |
1,233,526 | <train_model><EOS> | test[["SK_ID_CURR","TARGET"]].to_csv('xgb_submission_esi.csv', index=False, float_format='%.8f' ) | Home Credit Default Risk |
1,189,554 | <SOS> metric: AUC Kaggle data source: home-credit-default-risk<predict_on_test> | warnings.simplefilter(action='ignore', category=FutureWarning)
py.init_notebook_mode(connected=True)
%matplotlib inline
cf.go_offline()
| Home Credit Default Risk |
1,189,554 | prob=xgb.predict_proba(test_np )<train_model> | def one_hot_encoder(df, nan_as_category = True):
original_columns = list(df.columns)
categorical_columns = [col for col in df.columns if df[col].dtype == 'object']
df = pd.get_dummies(df, columns= categorical_columns, dummy_na= nan_as_category)
new_columns = [c for c in df.columns if c not in original_columns]
return... | Home Credit Default Risk |
1,189,554 | clf = RandomForestClassifier(n_estimators=10)
clf.fit(x_Train,y_Train[:,1])
score(clf, x_Test,y_Test[:,1] )<train_model> | num_rows = None
nan_as_category = True | Home Credit Default Risk |
1,189,554 | ada = AdaBoostClassifier(n_estimators=100, random_state=0)
ada.fit(x_Train,y_Train[:,1])
score(ada,x_Test,y_Test[:,1] )<predict_on_test> | print("Start Train Test................." ) | Home Credit Default Risk |
1,189,554 | prob3 = ada.predict_proba(test_np )<train_model> | df = pd.read_csv('.. /input/application_train.csv', nrows= num_rows)
test_df = pd.read_csv('.. /input/application_test.csv', nrows= num_rows)
print("Train samples: {}, test samples: {}".format(len(df), len(test_df)))
df = df.append(test_df ).reset_index()
del test_df
gc.collect() | Home Credit Default Risk |
1,189,554 | gbc = GradientBoostingClassifier(n_estimators=100, learning_rate=1.0,
max_depth=1, random_state=0 ).fit(x_Train, y_Train[:,1])
score(gbc, x_Test, y_Test[:,1])
<train_model> | for bin_feature in ['CODE_GENDER', 'FLAG_OWN_CAR', 'FLAG_OWN_REALTY']:
df[bin_feature], uniques = pd.factorize(df[bin_feature])
df, cat_cols = one_hot_encoder(df, nan_as_category)
df['DAYS_EMPLOYED'].replace(365243, np.nan, inplace= True ) | Home Credit Default Risk |
1,189,554 | rgl = LogisticRegression(random_state=0, solver='lbfgs',
multi_class='multinomial' ).fit(x_Train, y_Train[:,1])
score(rgl, x_Test, y_Test[:,1] )<predict_on_test> | df['DAYS_EMPLOYED_PERC'] = df['DAYS_EMPLOYED'] / df['DAYS_BIRTH']
df['INCOME_CREDIT_PERC'] = df['AMT_INCOME_TOTAL'] / df['AMT_CREDIT']
df['INCOME_PER_PERSON'] = df['AMT_INCOME_TOTAL'] / df['CNT_FAM_MEMBERS']
df['ANNUITY_INCOME_PERC'] = df['AMT_ANNUITY'] / df['AMT_INCOME_TOTAL'] | Home Credit Default Risk |
1,189,554 | prob4 = rgl.predict_proba(test_np )<train_model> | a = df['DAYS_EMPLOYED_PERC'].tolist()
a = [x for x in a if str(x)!= 'nan']
b = df['INCOME_CREDIT_PERC'].tolist()
b = [x for x in b if str(x)!= 'nan']
c = df['INCOME_PER_PERSON'].tolist()
c = [x for x in c if str(x)!= 'nan']
d = df['ANNUITY_INCOME_PERC'].tolist()
d = [x for x in d if str(x)!= 'nan'] | Home Credit Default Risk |
1,189,554 | BC_model = BaggingClassifier().fit(x_Train, y_Train[:,1])
score(BC_model, x_Test, y_Test[:,1] )<train_model> | print("End Train Test.................. " ) | Home Credit Default Risk |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.