使用 xgboost 进行二进制分类，得到 10000x2 yhat

Question

我是 xgboost 的新手，我正在尝试运行二元分类模型。

import pandas as pd

import xgboost as xgb
from sklearn.model_selection import train_test_split
from sklearn.metrics import roc_auc_score

from xgboost.sklearn import XGBClassifier


df_train = pd.read_csv('case2_training.csv')
df_test = pd.read_csv('case2_testing.csv')
df_train.shape, df_test.shape

# Exam the data
print(df_train.head())
print(df_test.head())

# Check for exsitence of null
df_train.info()

print('--------------------------------------------------------------')

X, y = df_train.iloc[:, :-1], df_train.iloc[:, -1]  # X = features Y = label

data_dmatrix = xgb.DMatrix(data=X, label=y)
# Split the data
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=123)


xgb = XGBClassifier(base_score=0.5, booster=None, colsample_bylevel=1,
                    colsample_bynode=1, colsample_bytree=0.5, gamma=0, gpu_id=-1,
                    importance_type='gain', interaction_constraints=None,
                    learning_rate=0.3, max_delta_step=0, max_depth=6,
                    min_child_weight=1,  monotone_constraints=None,
                    n_estimators=25, n_jobs=0, num_parallel_tree=1,
                    objective='binary:logistic', random_state=0, reg_alpha=0,
                    reg_lambda=1, scale_pos_weight=None, seed=0, subsample=0.5,
                    tree_method=None, validate_parameters=False, verbosity=None)
xgb.fit(X_train,y_train)

yhat = xgb.predict_proba(X_test)

print(yhat.shape)
print(yhat)
print(y_test.shape)
# train_accuracy = roc_auc_score(y_test, yhat)

这里的 yhat 给出了一个 (10000x2)，这不是我所期望的。它应该是 10000x1，然后我可以将它们四舍五入以与 y_test 进行比较。谁能帮忙？谢谢

Answer 1

您会得到每个数据点的给定 class 的概率。所以列中两个数字的总和 = 1，您可以通过取 np.argmax().

找到 class

使用 xgboost 进行二进制分类，得到 10000x2 yhat

Doing a binary classification using xgboost, getting a 10000x2 yhat

python

classification

xgboost