| 基于学习的证券市场专家预测意见合成研究 |
| |
| 引用本文： | 杨善林,朱卫东,任明仑. 基于学习的证券市场专家预测意见合成研究[J]. 系统工程学报, 2004, 19(1): 94-98 |
| |
| 作者姓名： | 杨善林 朱卫东 任明仑 |
| |
| 作者单位： | 合肥工业大学管理学院,安徽,合肥,230009 |
| |
| 基金项目： | 国家自然科学基金资助项目(70171033)，安徽省自然科学基金资助项目(00043607). |
| |
| 摘 要： | 针对证券市场的专家预测意见的特点，研究了将学习、证据理论与协同学理论引入证券市场专家预测意见合成的理论与方法．讨论了将专家的历史预测数据用基本可信数表示作为学习样本，对初选专家进行聚类分析选择有协同效应的专家．该方法用神经网络优化基本可信数的修正系数，再用Dempster合成规则进行预测意见的合成．实验结果表明该方法应用于解决实际问题时具有较好的效果．
|
| 关 键 词： | 证券市场 证据理论 神经网络 证券投资 专家预测意见合成 聚类分析 |
| 文章编号： | 1000-5781(2004)01-0094-05 |
Learning based combination of expert opinions in securities market forecasting |
| |
| YANG Shan-lin,ZHU Wei-dong,REN Ming-lun. Learning based combination of expert opinions in securities market forecasting[J]. Journal of Systems Engineering, 2004, 19(1): 94-98 |
| |
| Authors: | YANG Shan-lin ZHU Wei-dong REN Ming-lun |
| |
| Abstract: | According to the characteristics of experts forecasting in securities market, this paper presents a new theory and method for combination of expert opinions in securities market forecasting by introducing learning, theory of evidence and synergetics. this paper uses basic probability numbers as learning samples to represent historical experts forecast opinions, and selects collaborative experts through cluster analysis in candidate experts. A new method is developed to combine different forecast opinions, which modifies the basic probability numbers using a neural network and then adopts the Dempster combine rules in experts opinions combination. An empirical study shows its effect in solving application problems. |
| |
| Keywords: | learning neural networks theory of evidence synergetics expert opinion combination |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
