Abstract | ||
---|---|---|
This paper proposes a modified version of support vector machines (SVMs), called c -ascending support vector machines (c -ASVMs), to model nonstationary financial time series. c -ASVMs are obtained by a simple modification of the regularized risk function in SVMs whereby the recent epsilon-insensitive errors are penalized more heavily than the distant c -insensitive errors. This procedure is based on the prior knowledge that in the non-stationary financial time series, the recent past data could provide more important information than the distant past data. In the experiment, c -ASVMS are tested using three real futures collected from the Chicago Mercantile Market. It is shown that the c -ASVMs with the actually ordered sample data consistently forecast better than the standard SVMs, with the worst performance when the reversely ordered sample data are used. Furthermore, the c -ASVAfs use fewer support vectors than those of the standard SVMs', resulting in a sparser representation of solution. |
Year | DOI | Venue |
---|---|---|
2003 | 10.1109/CIFER.2003.1196277 | IEEE Conference on Computational Intelligence for Financial Engineering and Economics CIFEr |
Keywords | Field | DocType |
non-stationary,SVMs,structural risk minimization principle | Data mining,Learning automata,Futures contract,Computer science,Support vector machine,Financial time series forecasting,Artificial intelligence,Forecasting theory,Financial data processing,Risk function,Machine learning | Conference |
ISSN | Citations | PageRank |
2380-8454 | 3 | 0.45 |
References | Authors | |
6 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
juan cao li | 1 | 4 | 0.86 |
Kok Seng Chua | 2 | 59 | 5.78 |
Kian Guan Lim | 3 | 60 | 5.35 |
LJ Cao | 4 | 4 | 0.86 |
KS Chua | 5 | 4 | 0.86 |
LK Guan | 6 | 4 | 0.86 |