What does Precision@N mean?

[Borui Xu]

Dear author:

Thanks for your work! I read your paper but I have some questions about the evaluation metrics. I am confused about Precision@N. According to Appendix A.2., “The Precision@N is the proportion of top N predictions on each day with the positive label.” I don't understand what that means, maybe sometimes almost all stocks have negative labels, in addition I find the recall metrics are not satisfactory.

Looking forward to your reply!

[Wentao Xu]

Hi, the top N predictions are the N stocks with the largest prediction scores on each day (we sort the stocks on each day according to their prediction scores), and Precision@N is the proportion of top N predictions on each day with the positive label.

[Wentao Xu]

Yes, if all the labels of the top N prediction are positive, then the Precision@N is 1.
In the stock investment, since we do not know the stock future trends, we will invest the top N stocks with the largest prediction scores and hope that those N stocks would gain more positive returns.

[Borui Xu]

Thanks a lot!