xgboost
上进的菜鸟
这个作者很懒,什么都没留下…
展开
-
数据准备for xgboost
xgboost只接受数值1.分类的输出变量编码# multiclass classificationimport pandasimport xgboostfrom sklearn import model_selectionfrom sklearn.metrics import accuracy_scorefrom sklearn.preprocessing import La...翻译 2019-01-01 21:19:52 · 393 阅读 · 0 评论 -
存储xgboostmodel
# Train XGBoost model, save to file using pickle, load and make predictionsfrom numpy import loadtxtimport xgboostimport picklefrom sklearn import model_selectionfrom sklearn.metrics import accu...翻译 2019-01-01 21:22:39 · 522 阅读 · 0 评论 -
评估xgboost
1.划分训练测试集,看看测试集上好坏之前就有这里就不赘述了2.k-fold# k-fold cross validation evaluation of xgboost modelfrom numpy import loadtxtimport xgboostfrom sklearn.model_selection import KFoldfrom sklearn.model_...翻译 2019-01-01 21:36:04 · 881 阅读 · 0 评论 -
xgboost可视化
# plot decision treefrom numpy import loadtxtfrom xgboost import XGBClassifierfrom xgboost import plot_treeimport matplotlib.pyplot as plt# load datadataset = loadtxt('pima-indians-diabetes.csv...翻译 2019-01-01 21:40:27 · 7063 阅读 · 0 评论 -
变量重要性和变量选择in xgboost
1。变量重要得分# plot feature importance manuallyfrom numpy import loadtxtfrom xgboost import XGBClassifierfrom matplotlib import pyplot# load datadataset = loadtxt('pima-indians-diabetes.csv', delim...翻译 2019-01-01 21:50:59 · 7201 阅读 · 0 评论 -
early-stopping在xgboost的过拟合使用
https://machinelearningmastery.com/avoid-overfitting-by-early-stopping-with-xgboost-in-python/baseline,划分训练集测试集# monitor training performancefrom numpy import loadtxtfrom xgboost import XGBClas...翻译 2019-01-02 09:23:14 · 4752 阅读 · 0 评论 -
xgboost多线程
https://machinelearningmastery.com/best-tune-multithreading-support-xgboost-python/先占个坑翻译 2019-01-02 09:33:15 · 1342 阅读 · 0 评论 -
xgboost调节树的深度(size)和数量
1。树的个数(基分类器的个数# XGBoost on Otto dataset, Tune n_estimatorsfrom pandas import read_csvfrom xgboost import XGBClassifierfrom sklearn.model_selection import GridSearchCVfrom sklearn.model_selectio...翻译 2019-01-02 09:51:31 · 11007 阅读 · 0 评论 -
XGBOOST原理篇
mac安装conda install -c conda-forge xgboost个人理解:XGBOOST是先假设输的结构已经定了,求出来了最小的损失,然后根据该公式去分裂就得到第t次生成的树GBDT就是根据每轮样本与残差去按照普通的流程去建立CART回归树...原创 2019-02-26 20:53:25 · 232 阅读 · 0 评论