摘要
系统介绍了统计学习理论与支持向量机的基本思想,研究了它们在综合评价中的应用。分析了科研立项评审系统的设计方法,建立了基于的评审系统。文末比较了新评审系统和采用其它方法如模糊排序、神经网络等建立的评审系统所分别取得的拟合效果,比较结果SVM表明:采用支持向量机设计的评审系统结构简单、思路清晰且能取得更为理想的评审结果。
In this paper,the basic theory of statistical learning theory(SLT) and support vector machine(SVM)are introduced systemically and the application of SLT and SVM in the evaluation system are studied.For example, the design method of evaluation system for research projects supported by science foundation is analyzed.In order to improve the generalization performance,it reconstructs the evaluation system based on SVM.At the end of the paper,the comparison of the result of simulation between the evaluationsystem based on SVM and that based on other methods such as neural network and fuzzy sorting are discussed. The result shows that an ideal classification result can be acquired by adoption of SVM in this evaluation process. ;;
出处
《计算机工程》
CAS
CSCD
北大核心
2002年第8期28-30,共3页
Computer Engineering
基金
国家自然科学基金资助项目(70071023)
广东省自然科学基金资助项目(990828)
