TY  - RPRT
TI  - L1 regularization is better than L2 for learning and predicting chaotic systems
AU  - Z. Szabo
AU  - A. Lorincz
PY  - 2004
UR  - https://arxiv.org/abs/cs/0410015
ID  - cs/0410015
ER  -