Papers10:41Chinese content表示保留学习的残差代数这篇论文用残差代数替代特征拼接,在A股数据上收益从13.5%提升到19.1%,Sharpe从1.42涨到2.09,比所有对照方法都强。#FPRC-PQ#representation learning#quantitative investingaarXiv cs.LG@Yao WuSourceRead laterReadWorth followingUsefulFollow FPRC-PQ