ワイン品質分類

の精度を高める方法は?

予測モデル

を変更

独立変数の重みを削除または付与します(ワインの品質にはどのような重要な特性がありますか?)

スーパーパラメータ調整?

import pandas as pd
%matplotlib inline
import matplotlib.pyplot as plt
import seaborn as sns
import warnings
warnings.filterwarnings(action='ignore')

train = pd.read_csv('data/train.csv')
test = pd.read_csv('data/test.csv')

1.簡単なEDA

print(train.shape, test.shape)

(5497, 14) (1000, 13)

train.head(2)

index
quality
fixed acidity
volatile acidity
citric acid
residual sugar
chlorides
free sulfur dioxide
total sulfur dioxide
density
pH
sulphates
alcohol
type
0
0
5
5.6
0.695
0.06
6.8
0.042
9.0
84.0
0.99432
3.44
0.44
10.2
white
1
1
5
8.8
0.610
0.14
2.4
0.067
10.0
42.0
0.99690
3.19
0.59
9.5
red

test.head(2) # quality 변수가 없네.. quality가 target?

index
fixed acidity
volatile acidity
citric acid
residual sugar
chlorides
free sulfur dioxide
total sulfur dioxide
density
pH
sulphates
alcohol
type
0
0
9.0
0.31
0.48
6.6
0.043
11.0
73.0
0.9938
2.90
0.38
11.6
white
1
1
13.3
0.43
0.58
1.9
0.070
15.0
40.0
1.0004
3.06
0.49
9.0
red

train.info()  # 다행히 결측치는 없다

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 5497 entries, 0 to 5496
Data columns (total 14 columns):
 #   Column                Non-Null Count  Dtype
---  ------                --------------  -----
 0   index                 5497 non-null   int64
 1   quality               5497 non-null   int64
 2   fixed acidity         5497 non-null   float64
 3   volatile acidity      5497 non-null   float64
 4   citric acid           5497 non-null   float64
 5   residual sugar        5497 non-null   float64
 6   chlorides             5497 non-null   float64
 7   free sulfur dioxide   5497 non-null   float64
 8   total sulfur dioxide  5497 non-null   float64
 9   density               5497 non-null   float64
 10  pH                    5497 non-null   float64
 11  sulphates             5497 non-null   float64
 12  alcohol               5497 non-null   float64
 13  type                  5497 non-null   object
dtypes: float64(11), int64(2), object(1)
memory usage: 601.4+ KB

# train 데이터셋 변수간 상관관계(train) 보기 - heatmap
plt.figure(figsize=(12,12))
sns.heatmap(data = train.corr(), annot=True);

# train의 각 변수별 분포 확인 (subplot)
plt.figure(figsize=(12,12))
for i in range(1,13):
    plt.subplot(3,4,i)
    sns.distplot(train.iloc[:,i])
plt.tight_layout();
    # density 값은 왜 다르지?
    # 확률밀도함수의 y축이 density, 저 함수의 값을 적분하면 값이 1

# quality 변수를 기준 다른 피처들의 분포 확인 (barplot)
for i in range(11):
    fig = plt.figure(figsize = (12,6))
    sns.barplot(x= 'quality', y = train.columns[i+2], data = train)

    # 막대그래프는 평균? or 최빈값? , 가운데 선은 편차?
    # 오차막대(?) - 분산?

2.データの前処理

# type은 white와 red 두 종류인데 각각 0과 1로 변환
from sklearn.preprocessing import LabelEncoder

enc = LabelEncoder()
enc.fit(train['type'])
train['type'] = enc.transform(train['type'])
test['type'] = enc.transform(test['type'])

# 데이터 분리 및 불필요한 변수 제거
    # 독립변수 중 제거해도 좋은 변수가 있을까? 있다면 어떻게 알 수 있을까?
train_x = train.drop(['index','quality'], axis = 1)
train_y = train['quality']
test_x = test.drop('index', axis = 1)

train_x.shape, train_y.shape, test_x.shape

((5497, 12), (5497,), (1000, 12))

3.モデリング

# from sklearn.ensemble import RandomForestClassifier
# # 모델선언
# model = RandomForestClassifier()

# # 모델학습
# model.fit(train_x, train_y)

RandomForestClassifier()

from xgboost import XGBClassifier
from sklearn.metrics import roc_auc_score

# XGBClassifier 객체 생성
model = XGBClassifier(n_estimators=500, random_state=156) # 100 -> 500 (0.03 늘어남..)

# 학습 : 성능 평가 지표를 auc로 설정하고 학습 수행.
model.fit(train_x, train_y)

XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,
              colsample_bynode=1, colsample_bytree=1, enable_categorical=False,
              gamma=0, gpu_id=-1, importance_type=None,
              interaction_constraints='', learning_rate=0.300000012,
              max_delta_step=0, max_depth=6, min_child_weight=1, missing=nan,
              monotone_constraints='()', n_estimators=500, n_jobs=8,
              num_parallel_tree=1, objective='multi:softprob', predictor='auto',
              random_state=156, reg_alpha=0, reg_lambda=1,
              scale_pos_weight=None, subsample=1, tree_method='exact',
              validate_parameters=1, verbosity=None)

# 학습된 모델로 test 데이터 예측
y_pred = model.predict(test_x)

# 제출파일 생성
submission = pd.read_csv('data/sample_submission.csv')

submission['quality'] = y_pred

submission.to_csv('data/wine_quality_2.csv',index=False)