how to process beta factor in dl model

[yuxunwu]

I have some alpha factor libraries and beta factor libraries. The index storage of the alpha factor is two-dimensional (date, symbol), and the corresponding label is actually the rate of return of the symbol. My beta factor library is one-dimensional (date), and Different bata corresponds to different style labels, such as (size, mom,...), how should I handle the beta factor in the deep learning model, and add the beta factor to the model? I tried splicing the label of beta after the same date, but the effect of the model did not improve.

 

[yuxunwu]

plz，help me ，thanks very much ， I am using my alpha factor library to reproduce this article. The code of this article is open sourced in qlib. I found that the ic is as high as 8%. I hope to add the beta factor into it. The article is

Learning Multiple Stock Trading Patterns with Temporal Routing Adaptor and Optimal Transport​