TY - JOUR
T1 - Statistical mechanics of support vector networks
AU - Dietrich, Rainer
AU - Opper, Manfred
AU - Sompolinsky, Haim
PY - 1999
Y1 - 1999
N2 - Using methods of statistical physics, we investigate the generalization performance of support vector machines (SVMs), which have been recently introduced as a general alternative to neural networks. For nonlinear classification rules, the generalization error saturates on a plateau when the number of examples is too small to properly estimate the coefficients of the nonlinear part. When trained on simple rules, we find that SVMs overfit only weakly. The performance of SVMs is strongly enhanced when the distribution of the inputs has a gap in feature space.
AB - Using methods of statistical physics, we investigate the generalization performance of support vector machines (SVMs), which have been recently introduced as a general alternative to neural networks. For nonlinear classification rules, the generalization error saturates on a plateau when the number of examples is too small to properly estimate the coefficients of the nonlinear part. When trained on simple rules, we find that SVMs overfit only weakly. The performance of SVMs is strongly enhanced when the distribution of the inputs has a gap in feature space.
UR - http://www.scopus.com/inward/record.url?scp=0000149970&partnerID=8YFLogxK
U2 - 10.1103/PhysRevLett.82.2975
DO - 10.1103/PhysRevLett.82.2975
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
AN - SCOPUS:0000149970
SN - 0031-9007
VL - 82
SP - 2975
EP - 2978
JO - Physical Review Letters
JF - Physical Review Letters
IS - 14
ER -