提问人：zeman 提问时间：10/9/2023 更新时间：10/19/2023 访问量：43
如何在不删除列的情况下忽略 LGBMClassifier 中的 ID 列？
How to ignore ID column in LGBMClassifier without dropping column?
问：
我需要保留我的列来对测试数据进行预测，但模型用作预测变量，这是不需要的。感谢您的帮助！idid
x_train = train.drop(columns=['tgt'], axis=1)
y_train = train['tgt']
x_test = test_new.drop(columns=['tgt'], axis=1)
y_test = test_new['tgt']
rand = LGBMClassifier(seed=seed, learning_rate=0.1, num_leaves=12, n_estimators=700)
train_data = lgb.Dataset(x_train, label=y_train)
rand.fit(train_data)
答：
0赞
comphilip
10/19/2023
#1
尝试ignore_column参数。
rand = LGBMClassifier(seed=seed, learning_rate=0.1, num_leaves=12, n_estimators=700, ignore_column="name:tgt")
train_data = lgb.Dataset(train, label=y_train, ignore_column="name:tgt")
rand.fit(train_data)
0赞
PV8
10/19/2023
#2
您的代码中已经有了解决方案（几乎）：
x_train = train.drop(columns=['tgt', 'id'], axis=1)
y_train = train['tgt']
x_test = test_new.drop(columns=['tgt', 'id'], axis=1)
y_test = test_new['tgt']
rand = LGBMClassifier(seed=seed, learning_rate=0.1, num_leaves=12, n_estimators=700)
train_data = lgb.Dataset(x_train, label=y_train)
rand.fit(train_data)
评论