摘要
重学习是指从不只一个初始点开始的学习 ,本文分析了学习系统的重学习收敛性能 ,给出了描述重学习的结果优劣和初始点集稠密程度之间关系的定理 .并对一个优化问题引出的学习系统做了重学习的仿真 。
Multi learning is learning from more than one point called an initial point set. The convergence performance of multi learning is analysed, and a theorem describing the relationship between the qualification of the multi learning result and its initial points set's density is proved. As an example, an optimization problem is brought forward as a learning system and multi learnings of which from several initial points sets of different densities are simulated. The results demonstrate the relationship proposed by the theorem.
出处
《浙江大学学报（理学版）》
CAS
CSCD
2002年第2期213-218,共6页
Journal of Zhejiang University（Science Edition）
