Skip to content

300 features selection, CatBoost, NannyML to monitor concept drift, auto-retrain, daily load h5 trick

Notifications You must be signed in to change notification settings

GrigoriiTarasov/Optiver-Trading-at-the-Close-retrain-monitoring

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

51 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Optiver-Trading-at-the-Close-retrain-monitoring

Implementation of top solutions of Optiver - Trading at the Close 2023 with
ML engineering best practices extension

Stage Status
1 EDA w/ target as stock wap regression restore
2 Feature generation, 431 total
3 Revealed target handling
4 Batch learning on daily h5 data
5 Check aligment of inference and train features
6 Feature selection with Catboost
7 Params optimization with Optuna and MlFlow
on selected 300 features
8 Kaggle submission code
9 NannyMl to check perfomance decay for ts splits
10 Retrain options and DVC each fited porion

About

300 features selection, CatBoost, NannyML to monitor concept drift, auto-retrain, daily load h5 trick

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published