摘要
支持向量机的训练需要求解一个带约束的二次规划问题,但在数据规模很大情况下,经典训练方法将变得很困难。本文提出一种基于改进的混合蛙跳算法的SVM训练算法。针对混合蛙跳算法搜索速度慢且容易陷入局部极值的缺陷,将模拟退火思想引入到混合蛙跳算法中,提出一种改进的混合蛙跳算法。该算法保持了混合蛙跳算法参数少和容易实现的特点,同时通过模拟退火的降温过程来提高算法的进化速度和精度。实验结果表明,该算法能显著提高收敛速度,并能有效克服局部极值,在SVM训练中具有良好效果。
Since training SVM requires solving a restrained quadratic programming problem which becomes diffi- cult for large datasets, a improved Shuffled Frog Leaping Algorithm（SFLA） is proposed as an alternative to current algo- rithm. In order to overcome the defects of SFLA such as slow searching speed in evolution and local minimum, an im- proved algorithm in which the mechanism of Simulated Annealing（SA） is involved into basic SFLA is put forward in this paper. The proposed algorithm is almost as simple as SFLA and improves the evolution rate and precision through tem- perature decreasing procedures. The test results indicate that the algorithm enhances the convergence velocity outstand- ingly and averting the local extreme values effectively, and it is effective and feasible for SVM training.
出处
《信息化研究》
2011年第5期41-44,共4页
INFORMATIZATION RESEARCH
基金
国家自然科学基金项目(No:60872073No:51075068)
广东省自然科学基金(10252800001000001)资助项目
关键词
支持向量机
混合蛙跳算法
模拟退火
support vector machine（SVM）
shuffled frog leaping algorithm（SFLA）
simulated annealing（SA）
