Emsemble

RM

 # -*- coding: utf-8 -*-
 """
 RandomForestClassifier 예

 """
 import pandas as pd
 import numpy as np
 from sklearn.ensemble import RandomForestClassifier
 from sklearn.model_selection import train_test_split
 from sklearn.datasets import load_wine
 from sklearn import metrics #model 평가 도구

 #1.dataset load
 wine=load_wine()
 wine_x=wine.data
 wine_y=wine.target# 3개 범주 

 #data set보기
 print(wine_x[:5,:])
 """
 [[1.423e+01 1.710e+00 2.430e+00 1.560e+01 1.270e+02 2.800e+00 3.060e+00
   2.800e-01 2.290e+00 5.640e+00 1.040e+00 3.920e+00 1.065e+03]
  [1.320e+01 1.780e+00 2.140e+00 1.120e+01 1.000e+02 2.650e+00 2.760e+00
   2.600e-01 1.280e+00 4.380e+00 1.050e+00 3.400e+00 1.050e+03]
  [1.316e+01 2.360e+00 2.670e+00 1.860e+01 1.010e+02 2.800e+00 3.240e+00
   3.000e-01 2.810e+00 5.680e+00 1.030e+00 3.170e+00 1.185e+03]
  [1.437e+01 1.950e+00 2.500e+00 1.680e+01 1.130e+02 3.850e+00 3.490e+00
   2.400e-01 2.180e+00 7.800e+00 8.600e-01 3.450e+00 1.480e+03]
  [1.324e+01 2.590e+00 2.870e+00 2.100e+01 1.180e+02 2.800e+00 2.690e+00
   3.900e-01 1.820e+00 4.320e+00 1.040e+00 2.930e+00 7.350e+02]]
 """
 print(wine_y[:5]) #[0 0 0 0 0]
 print(wine_y[170:175]) #[2 2 2 2 2]

 #2. train/test
 X_train,X_test,y_train,y_test=train_test_split(wine_x,wine_y,test_size=0.3)

 #3.RM model 생성
 obj=RandomForestClassifier()
 model=obj.fit(X_train,y_train)
 print(model)
 """
 RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',
             max_depth=None, max_features='auto', max_leaf_nodes=None,
             min_impurity_decrease=0.0, min_impurity_split=None,
             min_samples_leaf=1, min_samples_split=2,
             min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=1,
             oob_score=False, random_state=None, verbose=0,
             warm_start=False)
 """

 #4.model 평가 : 도구
 pred=model.predict(X_test)
 Y=y_test

 # 평가 도구
 acc=metrics.accuracy_score(Y,pred)
 print(acc) #0.9629629629629629

 report=metrics.classification_report(Y,pred)
 print(report)
 """
              precision    recall  f1-score   support

           0       1.00      1.00      1.00        14
           1       0.90      1.00      0.95        18
           2       1.00      0.91      0.95        22

 avg / total       0.97      0.96      0.96        54
 """

 ##############################################
 # RF model Tuning
 ##############################################
 """
 n_estimators=10 tree개수 (400~500) 제일 좋음
 min_samples_split=2 : 변수의 개수(sqrt(n))
 """
 #2. train/test
 X_train,X_test,y_train,y_test=train_test_split(wine_x,wine_y,test_size=0.3)
 print(wine_x.shape)#(178, 13)  13의 루트
 print(np.sqrt(13)) #3.605551275463989=>4

 # 3. RM model 생성
 obj2=RandomForestClassifier(n_estimators=400,
                            min_samples_split=3)
 model2=obj2.fit(X_train,y_train)
 print(model2)
 """
 RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',
             max_depth=None, max_features='auto', max_leaf_nodes=None,
             min_impurity_decrease=0.0, min_impurity_split=None,
             min_samples_leaf=1, min_samples_split=4,
             min_weight_fraction_leaf=0.0, n_estimators=400, n_jobs=1,
             oob_score=False, random_state=None, verbose=0,
             warm_start=False)
 """
 # 4. model 평가 : 도구
 pred2=model.predict(X_test)
 Y=y_test

 # 평가 도구
 acc=metrics.accuracy_score(Y,pred2)
 print(acc) #0.9814814814814815

 report=metrics.classification_report(Y,pred2)
 print(report)
 """
              precision    recall  f1-score   support
           0       1.00      1.00      1.00        14
           1       0.95      1.00      0.97        18
           2       1.00      0.95      0.98        22
 avg / total       0.98      0.98      0.98        54
 """

cross_validation

 # -*- coding: utf-8 -*-
 """
 교차 검정예
 """
 import pandas as pd
 from sklearn.model_selection import cross_validate # 교차검정
 from sklearn.ensemble import RandomForestClassifier # RM

 # 1. data set
 iris=pd.read_csv("../data/iris.csv")
 print(iris.info())
 """
 <class 'pandas.core.frame.DataFrame'>
 RangeIndex: 150 entries, 0 to 149
 Data columns (total 5 columns):
 Sepal.Length    150 non-null float64
 Sepal.Width     150 non-null float64
 Petal.Length    150 non-null float64
 Petal.Width     150 non-null float64
 Species         150 non-null object
 dtypes: float64(4), object(1)
 memory usage: 5.9+ KB
 None
 """

 cols=list(iris.columns)

 x_data=iris[cols[:4]] #1~4
 y_data=iris[cols[-1]] 

 #2.model 생성
 obj=RandomForestClassifier()
 model=obj.fit(x_data,y_data)

 #3.교차 검정 cv=5(5겹 교차검정)
 score=cross_validate(model,x_data,y_data,cv=5)
 print(score)
 """
 {'fit_time': array([0.01000023, 0.01000023, 0.00900006, 0.00999999, 0.01000023]),
 'score_time': array([0.00099993, 0.00099993, 0.00099993, 0.00100017, 0.00099993]),
 'test_score': array([0.96666667, 0.96666667, 0.9       , 0.93333333, 1.        ]),
 'train_score': array([1., 1., 1., 1., 1.])}
 """

 test_score=list(score['test_score'])
 print(test_score) #[0.966, 0.966, 0.933, 0.9, 1.0]

 import numpy as np
 score_arr=np.array(test_score)
 print(score_arr.mean())#0.9533333333333334

RM_regression

 # -*- coding: utf-8 -*-
 """
 RandomForestRegressor 예
 """

 import pandas as pd
 import numpy as np

 from sklearn.ensemble import RandomForestRegressor
 from sklearn.model_selection import train_test_split
 from sklearn.datasets import load_boston # data set
 from sklearn import metrics # model 평가 도구

 # 1. dataset load
 boston = load_boston()
 boston_x = boston.data
 boston_y = boston.target # 연속형 

 # data set 보기
 print(boston_x.shape) # (506, 13)
 print(boston_y.shape) # (506,) 

 # 2. train/test
 x_train, x_test, y_train, y_test = train_test_split(
         boston_x, boston_y, test_size=0.3, random_state=123)

 # 3. RM model 생성
 obj = RandomForestRegressor(random_state=234)
 model = obj.fit(x_train, y_train)
 print(model)
 """
 RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=None,
            max_features='auto', max_leaf_nodes=None,
            min_impurity_decrease=0.0, min_impurity_split=None,
            min_samples_leaf=1, min_samples_split=2,
            min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=1,
            oob_score=False, random_state=234, verbose=0, warm_start=False)
 """

 # 4. model 평가 : 도구
 y_pred = model.predict(x_test)
 y_real_value = y_test 

 # 평가 도구
 model_score = model.score(x_test, y_test)
 print(model_score)
 # 0.7998066141697237

xgboost_test

 # -*- coding: utf-8 -*-
 """
 xgboot분류분석
 """
 import pandas as pd
 from xgboost import XGBClassifier #model
 from xgboost import plot_importance #중요변수 시각화
 from xgboost import plot_tree# tree 시각화
 from sklearn.model_selection import train_test_split

 #1.iris data set load
 iris=pd.read_csv("../data/iris.csv")

 cols=list(iris.columns)
 iris_x=iris[cols[:4]]
 iris_y=iris[cols[-1]]

 # 2. train/test set
 x_train,x_test,y_train,y_test=train_test_split(iris_x,iris_y,test_size=0.3,random_state=123)

 # 3. model 생성
 obj=XGBClassifier()
 model=obj.fit(x_train,y_train)
 print(model)
 """
 XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,
        colsample_bytree=1, gamma=0, learning_rate=0.1, max_delta_step=0,
        max_depth=3, min_child_weight=1, missing=None, n_estimators=100,
        n_jobs=1, nthread=None, objective='multi:softprob', random_state=0,
        reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,
        silent=True, subsample=1)

  objective = "binary:logistic": binary:logistic" : y변수 이항
 • max_depth = 2: tree 구조가 간단한 경우 : 2
 • nthread = 2 : cpu 사용 수 : 2
 • nrounds = 2 : 실제값과 예측값의 차이를 줄이기 위한 반복학습 횟수
 • eta = 1 : 학습률을 제어하는 변수(Default: 0.3), 오버 피팅을 방지

 """
 # 4. model 평가
 y_pred=model.predict(x_test)
 print(y_pred)
 Y=y_test
 """
 ['versicolor' 'virginica' 'virginica' 'versicolor' 'setosa' 'versicolor'
  'versicolor' 'setosa' 'setosa' 'versicolor' 'virginica' 'setosa'
  'versicolor' 'virginica' 'virginica' 'virginica' 'setosa' 'setosa'
  'versicolor' 'setosa' 'setosa' 'versicolor' 'setosa' 'virginica' 'setosa'
  'setosa' 'setosa' 'virginica' 'virginica' 'setosa' 'virginica'
  'versicolor' 'setosa' 'setosa' 'versicolor' 'versicolor' 'virginica'
  'setosa' 'setosa' 'versicolor' 'versicolor' 'setosa' 'virginica'
  'virginica' 'virginica']
 """

 # 중요변수 시각화
 import matplotlib.pyplot as plt
 plot_importance(model)
 plt.show()

 #fscore 중요변수 확인
 score=model.get_booster().get_fscore()
 print('x 중요변수=',score)
 #x 중요변수= {'Petal.Length': 255, 'Petal.Width': 135, 'Sepal.Width': 64, 'Sepal.Length': 118}

 #모델 평가
 from sklearn import metrics
 acc=metrics.accuracy_score(y_pred,Y)
 print("acc=",acc) #acc= 0.9333333333333333

 report=metrics.classification_report(Y,y_pred)
 print(report)
 """
              precision    recall  f1-score   support
      setosa       1.00      1.00      1.00        18
  versicolor       0.77      1.00      0.87        10
   virginica       1.00      0.82      0.90        17
 avg / total       0.95      0.93      0.93
 """

 plot_tree(model)
 plt.show()

xgboost_regression

 # -*- coding: utf-8 -*-
 """
 Created on Sun Feb 24 15:18:35 2019

 @author: 502-03
 """

 import pandas as pd
 from xgboost import XGBRegressor #model (회귀모델)
 from xgboost import plot_importance #중요변수 시각화
 from xgboost import plot_tree
 from sklearn.model_selection import train_test_split
 from sklearn.datasets import load_boston# dataset
 import matplotlib.pyplot as plt

 # 1. dataset load
 boston=load_boston()
 x=boston.data
 y=boston.target

 print(x.shape) #(506, 13)
 print(y.shape) #(506,)

 # 2. train/test set
 x_train,x_test,y_train,y_test=train_test_split(x,y,test_size=0.3,random_state=123)

 # 3. model 생성
 obj=XGBRegressor(n_estimators=400,max_depth=6)
 model=obj.fit(x_train,y_train)
 print(model)
 """
 XGBRegressor(base_score=0.5, booster='gbtree', colsample_bylevel=1,
        colsample_bytree=1, gamma=0, learning_rate=0.1, max_delta_step=0,
        max_depth=3, min_child_weight=1, missing=None, n_estimators=100,
        n_jobs=1, nthread=None, objective='reg:linear', random_state=0,
        reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,
        silent=True, subsample=1)
 """
 # 중요변수
 score=model.get_booster().get_fscore()
 print(score)
 """
 {'f5': 83, 'f12': 78, 'f0': 91, 'f4': 42, 'f7': 110, 'f10': 32,
 'f6': 46, 'f9': 38, 'f3': 1, 'f8': 16, 'f11': 51, 'f1': 2, 'f2': 15}
 """

 plot_importance(model)
 plt.show()

 plot_tree(model)
 plt.show()

 print(boston.feature_names)
 """
 ['CRIM' 'ZN' 'INDUS' 'CHAS' 'NOX' 'RM' 'AGE' 'DIS' 'RAD'
 'TAX' 'PTRATIO' 'B' 'LSTAT']
 """

xgboost_freeze

 # -*- coding: utf-8 -*-
 """
 수도사업소 주관 - big 콘테스트 dataset
 동파유무(0 or  1) 분류하는 위한 dataset
 """

 import pandas as pd
 from xgboost import XGBClassifier # model(분류모델)
 from xgboost import plot_importance # 중요변수 시각화
 from sklearn.datasets import load_boston # dataset
 from sklearn.model_selection import train_test_split

 from matplotlib import font_manager, rc
 font_name = font_manager.FontProperties(fname="c:/Windows/Fonts/malgun.ttf").get_name()
 rc('font', family=font_name)

 import matplotlib.pyplot as plt # 중요변수 시각화 

 freeze = pd.read_csv("../data/freeze_dataset.csv",encoding="MS949")
 print(freeze.info())
 '''
 RangeIndex: 37089 entries, 0 to 37088
 Data columns (total 95 columns):
 '''

 print(freeze.head())

 # 칼럼명 수정 : 공백 -> '_'
 freeze.columns = freeze.columns.str.replace(' ', '_')
 print(freeze.info())

 # 동파유무(0 or 1)
 print(freeze['동파유무'].value_counts())
 '''
 0.0    34130 : 90%
 1.0     2959 : 10%
 '''

 cols = list(freeze.columns) # 95개 칼럼
 x_cols = cols[1:]
 y_cols = cols[0]

 print(y_cols) # 동파유무

 train_set, test_set = train_test_split(
         freeze, test_size=0.4)

 # model
 obj = XGBClassifier()
 model = obj.fit(train_set[x_cols], train_set[y_cols])

 # 중요변수 score, 시각화
 score = model.get_booster().get_fscore()
 print(score)

 plot_importance(model)
 plt.show()