• 결정나무에서는 이런 전처리는 의미를 가지지 않는다.
• 주로 Cost Function을 최적화할 때 유효화할 때가 있다.
  
  
  
  
• Recall : 놓치면 안되는 데이터. ex) 암환자
  
  

와인 데이터 분석

데이터 전처리 전

import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt

red_url = 'https://raw.githubusercontent.com/PinkWink/ML_tutorial/master/dataset/winequality-red.csv'
white_url = 'https://raw.githubusercontent.com/PinkWink/ML_tutorial/master/dataset/winequality-white.csv'

red_wine = pd.read_csv(red_url, sep=';')
white_wine = pd.read_csv(white_url, sep=';')

red_wine['color']=1
white_wine['color']=0

wine = pd.concat([red_wine, white_wine])

import plotly.express as px

fig = px.histogram(wine, x='quality')
fig.show()

fig = px.histogram(wine, x='quality', color='color')
fig.show()

# 데이터 분리
from sklearn.model_selection import train_test_split
X = wine.drop(['color'], axis=1)
y = wine['color']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=13)

# 훈련데이터 와인 색 0,1 비율 확인
np.unique(y_train, return_counts=True)

# 학습
from sklearn.tree import DecisionTreeClassifier
from sklearn.metrics import accuracy_score

wine_tree = DecisionTreeClassifier(max_depth=2, random_state=13)
wine_tree.fit(X_train, y_train)

# 예측
y_pred_tr = wine_tree.predict(X_train)
y_pred_test = wine_tree.predict(X_test)

# 정확도
print(accuracy_score(y_train, y_pred_tr)) # 0.9553588608812776
print(accuracy_score(y_test, y_pred_test)) # 0.9569230769230769

데이터 전처리

import plotly.graph_objects as go

# 데이터 전처리 전
fig = go.Figure()
fig.add_trace(go.Box(y=X['fixed acidity'], name='fixed acidity'))
fig.add_trace(go.Box(y=X['chlorides'], name='chlorides'))
fig.add_trace(go.Box(y=X['quality'], name='quality'))
fig.show()

# MinMaxScaler, StandardScaler
from sklearn.preprocessing import MinMaxScaler, StandardScaler
mms = MinMaxScaler()
ss = StandardScaler()

X_mms = mms.fit_transform(X)
X_ss = ss.fit_transform(X)

X_mms_pd = pd.DataFrame(X_mms, columns=X.columns)
X_ss_pd = pd.DataFrame(X_ss, columns=X.columns)

# MinMaxScaler
fig = go.Figure()
fig.add_trace(go.Box(y=X_mms_pd['fixed acidity'], name='fixed acidity'))
fig.add_trace(go.Box(y=X_mms_pd['chlorides'], name='chlorides'))
fig.add_trace(go.Box(y=X_mms_pd['quality'], name='quality'))
fig.show()

# StandardScaler
fig = go.Figure()
fig.add_trace(go.Box(y=X_ss_pd['fixed acidity'], name='fixed acidity'))
fig.add_trace(go.Box(y=X_ss_pd['chlorides'], name='chlorides'))
fig.add_trace(go.Box(y=X_ss_pd['quality'], name='quality'))
fig.show()

학습 & 테스트

# MinMaxScaler
X_train, X_test, y_train, y_test = train_test_split(X_mms_pd, y, test_size=0.2, random_state=13)

wine_tree = DecisionTreeClassifier(max_depth=2, random_state=13)
wine_tree.fit(X_train, y_train)

y_pred_tr = wine_tree.predict(X_train)
y_pred_test = wine_tree.predict(X_test)

print(accuracy_score(y_train, y_pred_tr)) # 0.9553588608812776
print(accuracy_score(y_test, y_pred_test)) # 0.9569230769230769

# StandardScaler
X_train, X_test, y_train, y_test = train_test_split(X_ss_pd, y, test_size=0.2, random_state=13)

wine_tree = DecisionTreeClassifier(max_depth=2, random_state=13)
wine_tree.fit(X_train, y_train)

y_pred_tr = wine_tree.predict(X_train)
y_pred_test = wine_tree.predict(X_test)

print(accuracy_score(y_train, y_pred_tr)) # 0.9553588608812776
print(accuracy_score(y_test, y_pred_test)) # 0.9569230769230769

이진분류

wine['taste'] = [1. if grade>5 else 0. for grade in wine['quality']]

X = wine.drop(['taste', 'quality'], axis=1)
y = wine['taste']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=13)

wine_tree = DecisionTreeClassifier(max_depth=2, random_state=13)
wine_tree.fit(X_train, y_train)

y_pred_tr = wine_tree.predict(X_train)
y_pred_test = wine_tree.predict(X_test)

# 정확도
print(accuracy_score(y_train, y_pred_tr)) # 0.7294593034442948
print(accuracy_score(y_test, y_pred_test)) # 0.7161538461538461

# plot_tree
from sklearn.tree import plot_tree
plt.figure(figsize=(10,8))
plot_tree(wine_tree, feature_names=X.columns, rounded=True, filled=True);
plt.show()

Pipeline

from sklearn.pipeline import Pipeline

X = wine.drop(['color'], axis=1)
y = wine['color']

estimators = [('scaler', StandardScaler()),
              ('clf', DecisionTreeClassifier())]
pipe = Pipeline(estimators)

pipe.set_params(clf__max_depth=2)
pipe.set_params(clf__random_state=13)

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=13, stratify=y)

pipe.fit(X_train, y_train)

y_pred_tr = pipe.predict(X_train)
y_pred_test = pipe.predict(X_test)

print(accuracy_score(y_train, y_pred_tr)) # 0.9657494708485664
print(accuracy_score(y_test, y_pred_test)) # 0.9576923076923077

교차검증

# KFold
import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt

from sklearn.model_selection import train_test_split
from sklearn.model_selection import KFold
from sklearn.tree import DecisionTreeClassifier
from sklearn.metrics import accuracy_score

X = wine.drop(['color'], axis=1)
y = wine['color']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=13, stratify=y)

kfold = KFold(n_splits=5)
wine_tree_cv = DecisionTreeClassifier(max_depth=2, random_state=13)

cv_accuracy = []
for train_idx, test_idx in kfold.split(X):
    X_train, X_test = X.iloc[train_idx], X.iloc[test_idx]
    y_train, y_test = y.iloc[train_idx], y.iloc[test_idx]

    wine_tree_cv.fit(X_train, y_train)
    pred = wine_tree_cv.predict(X_test)
    cv_accuracy.append(accuracy_score(y_test, pred))

np.mean(cv_accuracy)  # 0.9142953751406407

# StratifiedKFold
from sklearn.model_selection import StratifiedKFold

skfold = StratifiedKFold(n_splits=5)
wine_tree_cv = DecisionTreeClassifier(max_depth=2, random_state=13)

cv_accuracy = []
for train_idx, test_idx in skfold.split(X, y):
    X_train, X_test = X.iloc[train_idx], X.iloc[test_idx]
    y_train, y_test = y.iloc[train_idx], y.iloc[test_idx]

    wine_tree_cv.fit(X_train, y_train)
    pred = wine_tree_cv.predict(X_test)
    cv_accuracy.append(accuracy_score(y_test, pred))

np.mean(cv_accuracy)  # 0.9604440101853496

# cross validation 간편히
from sklearn.model_selection import cross_val_score

skfold = StratifiedKFold(n_splits=5)
wine_tree_cv = DecisionTreeClassifier(max_depth=2, random_state=13)

cross_val_score(wine_tree_cv, X, y, scoring=None, cv=skfold)

# train score 같이 보기
from sklearn.model_selection import cross_validate

cross_validate(wine_tree_cv, X, y, scoring=None, cv=skfold, return_train_score=True)

GridSearchCV

from sklearn.model_selection import GridSearchCV

X = wine.drop(['color'], axis=1)
y = wine['color']

params = {'max_depth':[2,4,7,10]}
wine_tree = DecisionTreeClassifier(max_depth=2, random_state=13)

gridsearch = GridSearchCV(estimator=wine_tree, param_grid=params, cv=5)
gridsearch.fit(X, y)

# 결과 확인
import pprint

pp = pprint.PrettyPrinter(indent=4)
pp.pprint(gridsearch.cv_results_)

gridsearch.best_estimator_
gridsearch.best_score_
gridsearch.best_params_

# GridSearchCV, Pipeline
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler

X = wine.drop(['color'], axis=1)
y = wine['color']

estimators = [('scaler', StandardScaler()),
              ('clf', DecisionTreeClassifier())]
pipe = Pipeline(estimators)

param_grid = [{'clf__max_depth':[2,4,7,10]}]

gridsearch = GridSearchCV(estimator=pipe, param_grid=param_grid, cv=5)
gridsearch.fit(X, y)

print(gridsearch.best_estimator_)

# 데이터프레임으로 정리
score_df = pd.DataFrame(gridsearch.cv_results_)
score_df[['params', 'rank_test_score','mean_test_score','std_test_score']]

ROC 커브

# 수치 확인
from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, roc_auc_score, roc_curve

print('accuracy : ', accuracy_score(y_test, y_pred_test))
print('precision : ', precision_score(y_test, y_pred_test))
print('recall : ', recall_score(y_test, y_pred_test))
print('f1 : ', f1_score(y_test, y_pred_test))
print('auc : ', roc_auc_score(y_test, y_pred_test))

# accuracy :  0.9576923076923077
# precision :  0.9206349206349206
# recall :  0.90625
# f1 :  0.9133858267716536
# auc :  0.9403698979591837

# roc 커브 그리기
import matplotlib.pyplot as plt
%matplotlib inline

pred_proba = wine_tree.predict_proba(X_test)[:, 1]
fpr, tpr, thresholds = roc_curve(y_test, pred_proba)

plt.figure(figsize=(10,8))
plt.plot([0,1],[0,1])
plt.plot(fpr, tpr)
plt.title('ROC Curve')
plt.grid()
plt.show()