Have you experimented with other split values, 10% for training feels low to me.

…ime period specified (more on this in an article I will write on collecting & cleaning stock data). Then splits the most recent day of price data for test data (10%) whilst the other 90% is for training the machine learning model. We then run the algorithm specified above.
Algorithmic Stock Trading with XGBoost and Kalman Filters – Strategy
190
5
Sam Bennett
Jason Bell
·Follow
Mar 22, 2023
--
Have you experimented with other split values, 10% for training feels low to me. The numbers will show the proof I appreciate.
--
--
Written by Jason Bell
1.5K Followers
·149 Following
The Startup Quant and founder of ATXGV: Author of two machine learning books for Wiley Inc.
Responses (1)
Help
Status
About
Careers
Press
Blog
Privacy
Terms
Text to speech
Teams