Kaggle(캐글) Space Titanic Accuracy Score 81.21%

모듈 Import 작업

import pandas as pd

train = pd.read_csv('./input/train.csv')
test = pd.read_csv('./input/test.csv')
sample_submission = pd.read_csv('./input/sample_submission.csv')


import matplotlib.pyplot as plt
%matplotlib inline
import seaborn as sns
sns.set()
import warnings
warnings.filterwarnings("ignore")

시각화 도구

def bar_chart(feature):
    survived = train[train['Transported']==1][feature].value_counts()
    dead = train[train['Transported']==0][feature].value_counts()
    df = pd.DataFrame([survived,dead])
    df.index = ['Transported','Not Transported']
    df.plot(kind='bar',stacked=True, figsize=(10,5))

# 범위 설정 없으면 전체 범위에 대한 그래프
# 범위 설정 하면 해당 범위에 대한 그래프

def graph_chart(feature,min=0,max=0):
    facet = sns.FacetGrid(train, hue="Transported",aspect=4)
    facet.map(sns.kdeplot,feature,shade= True)
    facet.set(xlim=(0, train[feature].max()))
    facet.add_legend()
    
    if min == 0 and max == 0:
        plt.show()
    else:
        plt.xlim(min,max)

def scatter(feature):
    plt.figure(figsize=(15,7))
    sns.scatterplot(data=train,x=train.index,y=feature,hue='Transported',alpha=0.5)

Missing Value 처리

train_test = [train, test]
for dataset in train_test:
    for cols in dataset.columns:
        if((dataset[cols].dtype == 'float64') or (dataset[cols].dtype == 'int64')):
            dataset[cols].fillna(dataset[cols].mean(), inplace = True)
        else:
            dataset[cols].fillna(dataset[cols].mode()[0], inplace = True)

train.isnull().sum()

PassengerId     0
HomePlanet      0
CryoSleep       0
Cabin           0
Destination     0
Age             0
VIP             0
RoomService     0
FoodCourt       0
ShoppingMall    0
Spa             0
VRDeck          0
Name            0
Transported     0
dtype: int64

 

bool값 int로 바꿔주기 나머지 값들은 이후에 작업 진행

for dataset in train_test:
    dataset['VIP'] = dataset['VIP'].astype(int)
    dataset['CryoSleep'] = dataset['CryoSleep'].astype(int)

 

 

서비스에 대해서 하나라도 지불했는지 안했는지 여부 판단

for dataset in train_test:
    dataset['cost'] = dataset['RoomService'] + dataset['FoodCourt'] + dataset['ShoppingMall'] + dataset['Spa'] + dataset['VRDeck']

for dataset in train_test:
    dataset.loc[train['cost'] == 0, 'cost_clean'] = 0
    dataset.loc[train['cost'] > 0, 'cost_clean'] = 1

 

 

for dataset in train_test:
    dataset.loc[ dataset['Age'] <= 10, 'Age_clean'] = 0
    dataset.loc[(dataset['Age'] > 10) & (dataset['Age'] <= 16), 'Age_clean'] = 1
    dataset.loc[(dataset['Age'] > 16) & (dataset['Age'] <= 20), 'Age_clean'] = 2
    dataset.loc[(dataset['Age'] > 20) & (dataset['Age'] <= 26), 'Age_clean'] = 3
    dataset.loc[(dataset['Age'] > 26) & (dataset['Age'] <= 30), 'Age_clean'] = 4
    dataset.loc[(dataset['Age'] > 30) & (dataset['Age'] <= 36), 'Age_clean'] = 5
    dataset.loc[(dataset['Age'] > 36) & (dataset['Age'] <= 40), 'Age_clean'] = 6
    dataset.loc[(dataset['Age'] > 40) & (dataset['Age'] <= 46), 'Age_clean'] = 7
    dataset.loc[(dataset['Age'] > 46) & (dataset['Age'] <= 50), 'Age_clean'] = 8
    dataset.loc[(dataset['Age'] > 50) & (dataset['Age'] <= 60), 'Age_clean'] = 9
    dataset.loc[ dataset['Age'] > 60, 'Age_clean'] = 10

for dataset in train_test:
    dataset['Group'] = dataset['PassengerId'].str[:4]
    dataset['ID'] = dataset['PassengerId'].str[-2:]
    dataset['Group_Size']=dataset['Group'].map(lambda x: dataset['Group'].value_counts()[x])
    dataset.loc[dataset['Group_Size'] == 1, 'Solo'] = 1
    dataset.loc[dataset['Group_Size'] != 1, 'Solo'] = 0
    dataset['Group'] = dataset['Group'].astype('int')
    dataset['ID'] = dataset['ID'].astype('int')

HomePlanet_map = {"Earth": 0, "Europa": 1, "Mars": 2}

for dataset in train_test:
    dataset['HomePlanet_clean'] = dataset['HomePlanet'].map(HomePlanet_map)

Destination_map = {"TRAPPIST-1e": 0, "55 Cancri e": 1, "PSO J318.5-22": 2}

for dataset in train_test:
    dataset['Destination_clean'] = dataset['Destination'].map(Destination_map)

 

for dataset in train_test:
    dataset['Cabin_deck'] = dataset['Cabin'].apply(lambda x: x.split('/')[0])
    dataset['Cabin_number'] = dataset['Cabin'].apply(lambda x: x.split('/')[1]).astype(int)
    dataset['Cabin_side'] = dataset['Cabin'].apply(lambda x: x.split('/')[2])

Cabin_deck_map = {"A": 0, "B": 1, "C": 2, "D":3, "E":4, "F":5, "G":6, "T":7}

for dataset in train_test:
    dataset['Cabin_deck'] = dataset['Cabin_deck'].map(Cabin_deck_map)

Cabin_side_map = {"S": 0, "P": 1}

for dataset in train_test:
    dataset['Cabin_side'] = dataset['Cabin_side'].map(Cabin_side_map)

for dataset in train_test:
    dataset['Cabin_number'] = pd.qcut(dataset['Cabin_number'], 5)
    dataset['Cabin_number'] = dataset['Cabin_number'].astype('category').cat.codes

 

Cabin_deck

 

Cabin_side

 

Cabin_number

 

feature = [
    'ID',
    'CryoSleep',
    'Age_clean',
    'cost_clean',
    'RoomService',
    'FoodCourt',
    'ShoppingMall',
    'Spa',
    'VRDeck',
    'Group',
    'Group_Size',
    'Solo',
    'HomePlanet_clean',
    'Destination_clean',
    'Cabin_deck',
    'Cabin_number',
    'Cabin_side'
]

label = [
    'Transported',
]

data = train[feature]
target = train[label]

from sklearn.model_selection import train_test_split

# Lazypredict Library 
import lazypredict
from lazypredict.Supervised import LazyClassifier

#LightGBM library 
from lightgbm import LGBMClassifier 

# Sklearn Metrics 
from sklearn.metrics import classification_report , accuracy_score , confusion_matrix 
from IPython.display import clear_output

def splitted_data(features , label) : 
    x_train,x_test,y_train,y_test = train_test_split(features, label , test_size = 0.3)
    
    print(f"Shape of x_train : {x_train.shape}")
    print(f"Shape of y_train : {y_train.shape}")
    print(f"Shape of x_test : {x_test.shape}")
    print(f"Shape of y_test: {y_test.shape}")
    return x_train,x_test,y_train,y_test

x_train,x_test,y_train,y_test = splitted_data(data , target)

 

clf = LazyClassifier(verbose=0,
                     ignore_warnings=True,
                     custom_metric=None,
                     predictions=False,
                     random_state=12,
                     classifiers='all')

models, predictions = clf.fit(x_train , x_test , y_train , y_test)

models

lgbm_classifier = LGBMClassifier(random_state = 42 )
lgbm_classifier.fit(x_train,y_train)

y_pred = lgbm_classifier.predict(x_test)

score = accuracy_score(y_test , y_pred) 
percentage = score * 100 
print(f"Accuracy Score: {percentage:.2f}%")

Accuracy Score: 81.21%

 

xx_test = test[feature]
sample_submission['Transported'] = lgbm_classifier.predict(xx_test)
sample_submission.to_csv('6-submission.csv',index=False)