金融数据3

构建模型并评估

• 这里采用逻辑回归、svm和决策树；随机森林和XGBoost进行模型构建，采取的评价方式是accuracy

• 首先导入我们需要用的库

from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier 
from xgboost import XGBClassifier
from sklearn.model_selection import train_test_split

from sklearn.metrics import accuracy_score

X_train, X_test, y_train, y_test = train_test_split(df.drop('status', axis=1), df['status'], test_size=0.3, random_state=2018)

1. 逻辑回归

ltc = LogisticRegression()
ltc.fit(X_train, y_train)
predicted = ltc.predict(X_test)
accuracy_score(y_test, predicted)

0.7477224947442186

2. svm

svc = SVC()
svc.fit(X_train, y_train)
predicted = svc.predict(X_test)
accuracy_score(y_test, predicted)

0.7484232655921513

3. 决策树

dtc = DecisionTreeClassifier()
dtc.fit(X_train, y_train)
predicted = dtc.predict(X_test)
accuracy_score(y_test, predicted)

0.7484232655921513

4. 随机森林

rfc = RandomForestClassifier()
rfc.fit(X_train, y_train)
predicted = rfc.predict(X_test)
accuracy_score(y_test, predicted)

0.7687456201822004

5. XGBoost
   在利用XGBoost进行数据你和时，发现报错了，在检查特征类型时，发现“cust_id_bin”特征类型为"category"类型的，无法进行拟合，为了简单，我这里就直接删除了这个特征。。。

X_train.drop('cust_id_bin', axis=1, inplace=True)
X_test.drop('cust_id_bin', axis=1, inplace=True)

xgb = XGBClassifier()
xgb.fit(X_train, y_train)
xgb.predict(X_test)
accuracy_score(predicted, y_test)

0.7687456201822004