lightgbm调参后人工修改的结果记录20220216

sphinxrascal168

已于 2022-02-16 21:53:22 修改

阅读量650

点赞数

分类专栏：大幅度文章标签：机器学习 p2p 人工智能

于 2022-02-16 21:10:50 首次发布

本文链接：https://blog.youkuaiyun.com/sphinxrascal168/article/details/122971435

版权

大幅度专栏收录该内容

41 篇文章

订阅专栏

本文比较了LGBMClassifier的不同参数设置对模型性能的影响，包括正则化系数、叶子数、深度等，展示了不同配置下准确率的变化。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

761688

clf = LGBMClassifier(reg_alpha=0.15,reg_lambda=10,num_leaves=60,max_depth=7,n_estimators=1500,
learning_rate=0.01,bagging_fraction=0.7,bagging_freq=10,max_bin=150,
min_data_in_leaf=30,feature_fraction=0.7,random_state=1)

lgm
[[228612    373]
 [ 41600   2732]]
[[97835   202]
 [18095  1004]]

clf = LGBMClassifier(reg_alpha=0.5,reg_lambda=10,num_leaves=60,max_depth=7,n_estimators=1500,
learning_rate=0.01,bagging_fraction=0.7,bagging_freq=10,max_bin=150,
min_data_in_leaf=30,feature_fraction=0.7,random_state=1)

lgm
[[228605    380]
 [ 41581   2751]]
[[97832   205]
 [18076  1023]]

clf = LGBMClassifier(reg_alpha=0.5,reg_lambda=10,num_leaves=60,max_depth=7,n_estimators=1500,
learning_rate=0.01,bagging_fraction=0.7,bagging_freq=10,max_bin=100,
min_data_in_leaf=30,feature_fraction=0.7,random_state=1)

[[228618    367]
 [ 41581   2751]]
[[97838   199]
 [18084  1015]]

clf = LGBMClassifier(reg_alpha=0.5,reg_lambda=10,num_leaves=60,max_depth=7,n_estimators=1500,
learning_rate=0.01,bagging_fraction=0.6,bagging_freq=10,max_bin=100,
min_data_in_leaf=30,feature_fraction=0.7,random_state=1)

lgm
[[228623    362]
 [ 41643   2689]]
[[97826   211]
 [18094  1005]]