Portfolio Estimation Based on Support Vector Machine

摘要：

Traditional portfolio analysis supposes that the joint probability distribution function of yield vector is known, or uses the mean value and covariance of samples replacing that of joint distribution function. In this paper, the joint probability density function of yield vector is estimated based on support vector machine to realize the ability of generalization and anti-noise under a small quality of samples. Support vector machine is also used to estimate the optimised portfolio from samples directly. Experiments show that the results are satisfied.

关键词： Support vector machine Portfolio Joint probability density Function estimation

作者: Gensheng Hu

作者单位: Department of Computer Science Shangqiu Normal College Shangqiu, Henan Province,China

会议类型: 国际会议

会议名称: 2010 International Conference on Circuit and Signal Processing(2010年电路与信号处理国际会议 ICCSP 2010)

会议地点: 上海

会议语种:英文

页码: 282-285

在线出版日期: 2010-12-25（万方平台首次上网日期，不代表论文的发表时间）