使用scikit-learn的随机森林对西瓜进行分类

1、数据源：  http://blog.csdn.net/wiking__acm/article/details/50971461

2、参考： https://www.kaggle.com/jeffd23/titanic/scikit-learn-ml-from-start-to-finish/notebook中随机森林的使用

3、代码：

import pandas as pd

from sklearn import preprocessing

from sklearn.ensemble import RandomForestClassifier

from sklearn.model_selection import train_test_split

from sklearn.model_selection import GridSearchCV

from sklearn.metrics import accuracy_score,make_scorer

train_data = pd.read_csv('D:\\workspace\\kaggle\\data\\zhouzhihua-gua\\train_data.csv')

test_data = pd.read_csv('D:\\workspace\\kaggle\\data\\zhouzhihua-gua\\test_data.csv')

#将数据转化未label（0-N）形式

def encode_features(df_train, df_test):

      features = ['色泽', '根蒂', '敲声', '纹理', '脐部', '触感']

      df_combined = pd.concat([df_train[features], df_test[features]])

      

      for feature in features:

            le = preprocessing.LabelEncoder()

            le = le.fit(df_combined[feature])

            df_train[feature] = le.transform(df_train[feature])

            df_test[feature] = le.transform(df_test[feature])

      

      return df_train, df_test

 

def simplify_interval_info(df):

      bins_density = (0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8)

      bins_sugar = (0, 0.1, 0.2, 0.3, 0.4, 0.5)

      

      group_name_density = [0, 1, 2, 3, 4, 5, 6, 7]

      group_name_sugar = [0, 1, 2, 3, 4]

      

      category_density = pd.cut(df['密度'], bins_density, labels=group_name_density)

      categroy_sugar = pd.cut(df['含糖率'], bins_sugar, labels=group_name_sugar)

      

      df['密度'] = category_density

      df['含糖率'] = categroy_sugar

      

      return df

      

      

train_data, test_data = encode_features(train_data, test_data)

train_data = simplify_interval_info(train_data)

test_data = simplify_interval_info(test_data)

X_all = train_data.drop(['好瓜'], axis=1)

y_all = train_data['好瓜']

y_result = [1,0,0]

num_test = 0.50

X_train, X_test, y_train, y_test = train_test_split(X_all, y_all, test_size=num_test, random_state=3)

# Choose some parameter combinations to try

parameters = {'n_estimators':[5,6,7],

                     'criterion':['entropy', 'gini']

                     }

                     

# Type of scoring used to compare parameter combinations

acc_scorer = make_scorer(accuracy_score)

            

clf = RandomForestClassifier()

# Run the grid search

grid_obj = GridSearchCV(clf, parameters, scoring=acc_scorer)

grid_obj = grid_obj.fit(X_train, y_train)

# Set the clf to the best combination of parameters

clf = grid_obj.best_estimator_

clf = clf.fit(X_train, y_train)

test_predictions = clf.predict(X_test)

print("测试集准确率:   %s " % accuracy_score(y_test, test_predictions))

predictions = clf.predict(test_data)

print("最终准确率:   %s " % accuracy_score(y_result, predictions))