DocumentCode :
809132
Title :
Lower Bounds for the Empirical Minimization Algorithm
Author :
Mendelson, Shahar
Author_Institution :
Centre for Math. & its Applic., Australian Nat. Univ., Canberra, ACT
Volume :
54
Issue :
8
fYear :
2008
Firstpage :
3797
Lastpage :
3803
Abstract :
In this correspondence, we present a simple argument that proves that under mild geometric assumptions on the class F and the set of target functions T, the empirical minimization algorithm cannot yield a uniform error rate that is faster than 1/radic(k) in the function learning setup. This result holds for various loss functionals and the target functions from T that cause the slow uniform error rate are clearly exhibited.
Keywords :
error statistics; function approximation; learning (artificial intelligence); minimisation; statistical distributions; empirical minimization algorithm; function approximation; function learning problem; lower bound; probability distribution; statistical learning theory; target function set; uniform error rate; Australia Council; Cities and towns; Error analysis; Mathematics; Minimization methods; Pattern recognition; Performance loss; Probability distribution; Random variables; Statistical learning; Empirical minimization; function learning; lower bounds; statistical learning theory;
fLanguage :
English
Journal_Title :
Information Theory, IEEE Transactions on
Publisher :
ieee
ISSN :
0018-9448
Type :
jour
DOI :
10.1109/TIT.2008.926323
Filename :
4567590
Link To Document :
بازگشت