hyperparameter tuning

[pwner4810]

hey sir thanks for your helpful documentation I have a question and i was wondering if yo could answer have you tried hyperparameter tuning like bayesian optimization and why you havent used this in your project

thanks

[BruceYanghy]

hey sir thanks for your helpful documentation I have a question and i was wondering if yo could answer have you tried hyperparameter tuning like bayesian optimization and why you havent used this in your project

thanks

I tried Grid Search and Random Search. What I didn't do in this paper is: tuning the parameters along the way, I only used the in-sample data from 2009-2015 to tune the parameters once, so there is some alpha decay here. Ideally, you should tune the parameters every quarter when you retrain your model.

Numerous hyperparameters – e.g. the learning rate, total_timesteps – influence the learning process and are usually determined by testing some variations.