Quadratic bienenstock-cooper-munro radial basis fuction network intellient stock trading
As self-reorganizing learning approaches develops over the years under time-variant conditions, these mechanisms need to reorganize fuzzy-associative knowledge in real-time dynamic environment to maximize their operating values. Financial houses are relying heavily on these systems to address real w...
Saved in:
| 主要作者: | |
|---|---|
| 其他作者: | |
| 格式: | Final Year Project |
| 語言: | English |
| 出版: |
2011
|
| 主題: | |
| 在線閱讀: | http://hdl.handle.net/10356/45292 |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 機構: | Nanyang Technological University |
| 語言: | English |
| 總結: | As self-reorganizing learning approaches develops over the years under time-variant conditions, these mechanisms need to reorganize fuzzy-associative knowledge in real-time dynamic environment to maximize their operating values. Financial houses are relying heavily on these systems to address real world complex systems in their trading operations. Although Hebbian theory is the basic computational framework for associative learning, it is unfavourable for time-variant online-learning as it suffers from stability limitation and impedes unlearning. Therefore, QBCM is adopted because of its neurological learning via meta-plasticity principles that provides associative and dissociative learning. This project focuses on the interpretation of QBCM theory for a self-organising learning system based on RBF. From the experimental results, the analysis of FTSB MIB index time series and chaotic time series by Lorenz using QBCM-RBF and RBF network showed that QBCM-RBF is able to forecast a better prediction on the amount of rise and fall with smaller errors. Further enhancements are done with moving averages and signals of indications provided are reasonably well for consideration. |
|---|