more test, more hope
You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
lj020 3a8eae6e5c
update readme.md
4 years ago
.vscode normal update, not many things 4 years ago
catboost_info add many things 4 years ago
codes normal update, not many things 4 years ago
create_datas normal update, not many things 4 years ago
nn_LYC add many things 4 years ago
nn_TR add many things 4 years ago
nn_ZSH add many things 4 years ago
regressors add many things 4 years ago
traindata normal update, not many things 4 years ago
README.md update readme.md 4 years ago
TLE_to_train.py add many things 4 years ago
TLEdata_extract.py add many things 4 years ago
test_full.py add many things 4 years ago

README.md

ssa_everyone

more test, more hope

目前更新的regressor技术都在同名文件夹使用有点怪因为调整都在对应python里而不是总的test里。

how to use

test_full.py可以试试就在这master位置打开把要不然感觉可能会混乱位置

details

folder info
codes 存储的是制造数据集的代码通过处理过的parquet文件
create_datas 存储的是制造出来的数据集文件名包括seed信息同parquet和同seed创建相同
linear 存储线性模型
models 存储回归(决策树)模型
nn_models 存储nn模型
nn_XXX (3个) 存储xxx写的源代码文件
regressors 存储各种回归预测方式代码
traindata 原初下载的数据文件TLE处理后

Regressors files details

  • linear_study线性学习加权
  • linear_sum_regressor线性加权设置权值
  • nn_usenn使用内容定义class
  • normal_use通用一些的import和求averageR2
  • one_nn对单个nn模型的使用
  • one_regressor对单个决策树模型的使用
  • stacking_regressorstacking方案调整方案比较多

out-of-data

然后新regressor就往这里丢了终于是分开了天——

细节可能需要修一修average也有了seed可选model太大就不传勒