Agent-Based Portfolio Selection with Data Mining Ability
Portfolio selection is about how to determine a most suitable portfolio for the large private or institutional investor. It is a key step in financial investment. There are many portfolio selection models available. Each model has its own pros and cons. Therefore, we need to combine these models in real applications. In this paper, we propose a multi-agent architecture for this purpose. Such an architecture is very adaptive and robust. Before using different models to select a final portfolio for an investor, we have to construct a small set of promising securities from a large set of securities, which is based on historical fundamental and technical data about securities. It is infeasible to manually analysis large sets of data efficiently. Thus we introduce data mining techniques and incorporate them into the proposed multi-agent system to perform this task. Put all these together, the proposed approach is a very useful strategy for improving the performance of portfolio selection and facilitating the construction of other hybrid intelligent systems.
Zili Zhang Chengqi Zhang
School of Computing and Mathematics Faculty of Science and Technology Deakin University Geelong Victoria 3217, Australia
国际会议
8th International Conference on Neural Information Processing(ICONIP 2001)(第八届国际神经信息处理大会)
上海
英文
591-596
2001-11-14（万方平台首次上网日期，不代表论文的发表时间）