python|LightGBM模型_lgb.lgbmclassifier

# -*- coding: utf-8 -*-
"""
Created on Fri Jun 12 16:20:17 2020
@author: weiping
"""
 
import xgboost as xgb
import lightgbm as lgb  
from sklearn.model_selection  import train_test_split
from sklearn.metrics import *
from sklearn.datasets import load_iris
 
iris = load_iris()
 
data_x = iris.data 
data_y = iris.target
 
x_tr,x_te,y_tr,y_te = train_test_split(data_x,data_y,train_size = 0.7,random_state =22)
#XGBboost模型
xgb_model = xgb.XGBClassifier()
xgb_model.fit(x_tr,y_tr)
xgb_predict = xgb_model.predict(x_te)
 
print("xgb准确率：" ,str(xgb_model.score(x_te,y_te)))
#print("roc_auc_score:",str(roc_auc_score(y_te,xgb_predict))) 不支持多分类
print("precision_score:",str(precision_score(y_te,xgb_predict,average = 'weighted')))
print("recall_score:" , str(recall_score(y_te,xgb_predict,average = 'weighted')))
print("f1_score:",str(f1_score(y_te,xgb_predict,average = 'weighted')))
'''
xgb准确率： 0.9333333333333333
precision_score: 0.9344662309368191
recall_score: 0.9333333333333333
f1_score: 0.9332681655262302
'''
 
#LGB模型
'''
模型参数
objective: (objective_type, app, application)
回归任务：
'regression'（默认）
'poisson'
'tweedie'
分类任务：
'binary'：二分类
'multiclass'：多分类
boosting_type:(boosting_type, boost)
'gbdt'：传统的梯度提升决策树 （默认）
'rf'：随机森林
'dart'：Dropouts meet Multiple Additive Regression Trees
'goss'：Gradient-based One-Side Sampling 训练更快，可能欠拟合
data：训练数据
valid：验证数据集
num_iteration:(num_trees, n_estimators)
迭代次数，100（默认）
learning_rate:(eta, shrinkage_rate)
衰减因子，0.1（默认）
seed:(random_state, random_seed)
num_threads:(n_jobs, nthreads)
num_leaves:
单棵树的最大叶子数，31（默认）
控制学习参数：
'max_depth':
树的最大深度，-1（默认），无限制
可用于控制过拟合
'lambda_l2':(reg_lambda, lambda)
L2正则化，0（默认）
'lambda_l1':(reg_alpha)
L1正则化，0（默认）
'min_data_in_leaf'(min_data, min_child_samples)
一个叶子的最小数据量，20（默认）
可用于控制过拟合
'subsample': (sub_row, bagging, bagging_fraction)
对样本进行采样，1（默认）
可用于控制过拟合
'sub_feature': (colsample_bytree, feature_fraction)
对特征进行采样，1（默认）
加速训练，控制过拟合
early_stopping: (early_stopping_round)
'''
 
lgb_model = lgb.LGBMClassifier()
lgb_model.fit(x_tr,y_tr)
lgb_predict = lgb_model.predict(x_te)
 
 
print("lgb准确率：" ,str(lgb_model.score(x_te,y_te)))
#print("roc_auc_score:",str(roc_auc_score(y_te,xgb_predict))) 不支持多分类
print("precision_score:",str(precision_score(y_te,lgb_predict,average = 'weighted')))
print("recall_score:" , str(recall_score(y_te,lgb_predict,average = 'weighted')))
print("f1_score:",str(f1_score(y_te,lgb_predict,average = 'weighted')))
'''
lgb准确率： 0.9555555555555556
precision_score: 0.9555555555555556
recall_score: 0.9555555555555556
f1_score: 0.9555555555555556
'''