L1 regularization is better than L2 for learning and predicting chaotic systems

Simple item page
dc.creatorSzabo, Z.
dc.creatorLorincz, A.
dc.date2004-10-07
dc.date.accessioned2026-07-07T03:21:50Z
dc.date.available2026-07-07T03:21:50Z
dc.descriptionEmergent behaviors are in the focus of recent research interest. It is then of considerable importance to investigate what optimizations suit the learning and prediction of chaotic systems, the putative candidates for emergence. We have compared L1 and L2 regularizations on predicting chaotic time series using linear recurrent neural networks. The internal representation and the weights of the networks were optimized in a unifying framework. Computational tests on different problems indicate considerable advantages for the L1 regularization: It had considerably better learning time and better interpolating capabilities. We shall argue that optimization viewed as a maximum likelihood estimation justifies our results, because L1 regularization fits heavy-tailed distributions -- an apparently general feature of emergent systems -- better.
dc.description13 pages, 4 figures
dc.identifierhttps://arxiv.org/abs/cs/0410015
dc.identifierhttp://arxiv.org/abs/cs/0410015
dc.identifier.urihttp://salesiana.dossiersoluciones.com/handle/123456789/32360
dc.subjectMachine Learning
dc.subjectArtificial Intelligence
dc.titleL1 regularization is better than L2 for learning and predicting chaotic systems
dc.typetext

Files

Collections

arXiv