A package to win data competition
Project description
lianyhaii
usage
第一步:安装与测试
pip install lianyhaii
import lianyhaii
print(lianyhaii.__version__)
第二步:准备数据、特征、参数
import pandas as pd
import numpy as np
import lianyhaii
import warnings
import sys
warnings.filterwarnings('ignore')
## 定义数据集、label、训练参数、特征名
train = pd.DataFrame({
'y':(np.random.random(size=500)>0.5).astype(int)
})
test = pd.DataFrame({
'y':(np.random.random(size=500)>0.5).astype(int)
})
for i in range(10):
train[f'x{i}'] = np.random.random(size=500)
test[f'x{i}'] = np.random.random(size=500)
label = 'y'
lgb_params = {
'objective': 'binary',
'boosting_type': 'gbdt',
'metric': 'auc',
'early_stopping_rounds': 50,
'verbose':-1,
}
base_features = [f'x{i}' for i in range(10)]
以上是送进模型的主要准备工作,接下来会轻松不少
第三步:快速模型测试
mt = make_test(train,test,base_features=base_features,new_features=[],
m_score=[[0.0,]],label=label,metrices=['auc'],log_tool=None)
mt.init_CV(seed=412,CV_type='skFold',n_split=5)
oof,pred = mt.lgb_test(lgb_params=lgb_params)
## 得到oof和pred方便后续调整或者提交
获奖经历
2021 山东赛 公积金贷款逾期预测 A榜 rank2/xxx B榜 rank6/xxx
2021 梧桐杯 5G潜客识别 B榜 rank7/xxx
2022 招商银行Fintech rank23/xxx -> 开源 2022 搜狐校园 情感分析 × 推荐排序 算法大赛 rank5/xxx -> 开源
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
lianyhaii-0.0.143.tar.gz
(126.9 kB
view hashes)
Built Distribution
lianyhaii-0.0.143-py3-none-any.whl
(115.4 kB
view hashes)
Close
Hashes for lianyhaii-0.0.143-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | d593d2f315319b838e65769bacc97d871fecdefeeb122dd09e95ab71fa98d1f9 |
|
MD5 | 97f912b47afed429d28f85d93375d6ef |
|
BLAKE2b-256 | ccab041c10a3bd5db41e47568968a74aaf8b5a7d28086d065a8802b33bb01548 |